# Full Text: Template Madlib: Deterministic Token Injection for Conditional IMRAD Manuscripts

> Extracted from `Friedman_2026_Template_d9248f4f.pdf`

---

## Page 1

Template Madlib: Deterministic Token Injection
for Conditional IMRAD Manuscripts
A public exemplar for source-owned lexical composition
Daniel Ari Friedman
Active Inference Institute
daniel@activeinference.institute
ORCID: 0000-0001-6232-9096
DOI: 10.5281/zenodo.20786638
2026-06-26

## Page 2

Contents
1
Abstract
2
2
Introduction: Lexicon as Data and Manuscript as Build Artifact
3
2.1
Contribution Ledger . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3
3
Methods: Source-Owned Token Injection and Conditional IMRAD Assembly
4
3.1
Design Principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5
3.2
Operational Phases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6
3.3
Protocol Steps
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
7
3.4
Section Plan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
9
3.5
Audit Rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11
4
Results: Provenance, Density, and Resolved Manuscript Surface
13
4.1
Token Inventory
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
13
4.2
Provenance Matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
14
5
Discussion: Accountability Boundaries for Generated Prose
18
6
Configuration: Schema-Controlled Lexicon, Slots, and Narrative Moves
19
6.1
Declared Section Titles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
19
6.2
Configuration Counts
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
20
6.3
Configured Field Summary
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
20
6.4
Configured Field Inventory
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
20
7
Evaluation: Gate Criteria, QA Probes, and Failure Discovery
26
7.1
Evaluation Criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
26
7.2
Quality Probes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
27
8
Reproducibility: Seeded Regeneration and Artifact Trace
29
9
Limitations: Non-Claims, Misuse Modes, and Human Review
30
9.1
Failure Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
30
10 Scope: Related Generators and Responsible Forking
32
11 Authoring Contract: Human Review and Forking Obligations
33
11.1 Authoring Obligations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
33

## Page 3

1
Abstract
This exemplar asks whether a reviewable pipeline can hydrate a complete IMRAD manuscript from configuration-owned lexical data
while preserving an audit trail that remains readable before and after rendering. The project deliberately keeps playful Mad Lib
mechanics inside a serious reproducibility contract: the manuscript shell names large placeholders, the config declares allowable
language, and the source code decides what text is emitted.
The committed seed is 431 and the current schema expands 22 slot rule(s) into 40 token choice(s) across 10 lexicon categories. The
configured narrative moves are state the manuscript-generation problem, name the deterministic intervention, and summarize the
audit surface. The central hypothesis is: Deterministic lexical injection can generate a complete conditional IMRAD manuscript
while preserving token provenance, section intent, and audit-ready method evidence.
The result is not a claim that lexical substitution creates scholarship. It is a worked template for conditional manuscript assembly:
section enablement, token provenance, figure registration, and unresolved-placeholder checks all become inspectable artifacts before
the shared renderer produces PDF, HTML, and slides.

## Page 4

2
Introduction: Lexicon as Data and Manuscript as Build Artifact
Mad Lib style generation is usually treated as a toy because it foregrounds the visible blank rather than the source of the replacement.
In a research pipeline, the blank is not the hard part. The hard part is making every replacement reviewable, deterministic, and
honest about what it can support. template_madlib turns that constraint into the subject of the exemplar.
The project treats nouns such as pipeline, protocol, section, lexicon and verbs such as hydrate, condition, bind, bind as versioned
data. Changing a lexicon entry is therefore closer to changing an input table than editing prose in place. That distinction matters
because the generated manuscript can be rerendered, diffed, and validated without asking the reader to trust an invisible drafting
session.
The introduction is configured to separate playful Mad Lib syntax from research claims, identify drift between prose and source data,
frame configuration as an inspectable dataset, and position conditional prose as a reproducibility problem. Those moves keep the
manuscript from pretending to be an open-ended language model. It is a bounded template: authors declare categories, slots, section
switches, method steps, and claim boundaries; source code transforms those declarations into manuscript bodies and evidence tables.
This exemplar is useful for protocols, educational scaffolds, review forms, templated reports, and other documents where conditional
text is unavoidable but should never become untraceable. The same pattern can be extended with domain-specific validators while
leaving the shared rendering infrastructure untouched.
2.1
Contribution Ledger
Claim
Boundary
A Mad Lib manuscript can remain reproducible when the
lexicon is treated as data.
Local exemplar claim; no live DOI or standalone release implied.
Conditional IMRAD section bodies can be rendered without
shared renderer changes.
Local exemplar claim; no live DOI or standalone release implied.
Large-grain manuscript variables can preserve author-readable
source files while still producing a complete manuscript.
Local exemplar claim; no live DOI or standalone release implied.
Token provenance can connect playful lexical substitution to
serious publication hygiene.
Local exemplar claim; no live DOI or standalone release implied.
A generated Methods section can be methodologically useful
when protocol rows, phases, figures, and validation gates share
one config-owned source.
Local exemplar claim; no live DOI or standalone release implied.
Configured-field origin tracking can make loader defaults visible
enough for reviewers and downstream forks.
Local exemplar claim; no live DOI or standalone release implied.
Pipeline-owned output regeneration can keep PDF, HTML,
slides, data, reports, and copied deliverables aligned without
hand-editing generated artifacts.
Local exemplar claim; no live DOI or standalone release implied.
A generated-method exemplar can make review handoff
auditable when the review packet includes source config, data
artifacts, figures, validation results, and copy statistics.
Local exemplar claim; no live DOI or standalone release implied.
Fork migration guidance can reduce overclaiming when it names
the source, test, validator, and evidence surfaces that must
change before domain use.
Local exemplar claim; no live DOI or standalone release implied.

## Page 5

3
Methods: Source-Owned Token Injection and Conditional IMRAD Assembly
The method is conditional section hydration. Each slot combines the seed, slot name, category, ordinal, and full category list into a
SHA-256 digest; the digest indexes the configured category. Including the full category list in the digest input means that a lexicon
edit can change the plan in a deterministic and reviewable way instead of silently preserving stale output.
The deterministic digest recipe is deliberately narrow. It does not sample from ambient prose, project history, or renderer state;
it uses only the committed seed, the slot declaration, the category name, the ordinal for repeated slots, and the ordered category
inventory. A fork can therefore explain a changed token choice as a changed seed, slot, or lexicon row rather than as an opaque
generation event.
The first review scenario is declared before generation. The project names its scope as a local exemplar, records enabled sections,
keeps DOI and publication claims blank, and treats the copy stage as a review handoff. That ordering matters because a reader
should know the allowed claim boundary before inspecting fluent generated text.
The governing constraint is publication claims stay local until release. The source manuscript is intentionally sparse: it contains
section titles and large-grain placeholders, not generated claims. The project script first validates the madlib: block, expands slots
into a TokenPlan, builds section bodies, writes artifact JSON, emits a figure registry, and only then writes hydrated Markdown under
output/manuscript/.
The configured method moves are validate config before composition, declare review scenario and method invariants, separate explicit
YAML fields from loader defaults, govern lexicon categories as reviewable data, expand slots through seeded digest selection, allocate
slot choices to enabled manuscript sections, compose conditional section bodies, assemble method evidence tables and visual audit
figures, align generated claims with the claim ledger, emit evidence artifacts before rendering, verify that no unresolved placeholders
remain, assemble a reviewer packet from regenerated artifacts, and preserve human review before publication claims. The protocol
sequence is Ingest declared manuscript schema produces MadlibConfig; Declare review scenario produces review scenario; Track field
origin produces explicit/default path inventory; Govern lexicon categories produces validated lexicon; Construct digest selection
material produces digest input records; Record selection invariants produces selection invariant set; Expand slot declarations
produces TokenPlan; Apply section conditions produces enabled section set; Compose conditional IMRAD bodies produces section
variables; Assemble evidence tables produces Markdown evidence tables; Align claims with evidence ledger produces claim-aligned
evidence surface; Generate visual audit surface produces registered figure set; Emit machine-readable artifacts produces output/
data, output/reports, and output/figures; Hydrate manuscript shell produces output/manuscript; Render and validate deliverables
produces validated project output; Assemble reviewer packet produces review packet; Copy review surface produces copied public
ation-review bundle; Document fork migration produces fork migration notes. These steps make the method auditable from three
directions: tests inspect the Python behavior, generated artifacts expose the token plan, and manuscript validation confirms that no
unresolved placeholder survives into rendered outputs.
The config-origin inventory currently separates 125 explicit YAML path(s) from 11 loader-defaulted path(s).
Treating origin as
method evidence prevents a rendered field from looking equally authored when it was actually inherited from the loader. The same
inventory drives configured-field tables and figures, so reviewers can inspect which schema blocks were intentionally set before judging
the generated prose.
Method invariants are reviewed as their own artifact. Token choices are allowed to change when the seed, slot name, category, ordinal,
or ordered category inventory changes; they are not allowed to change because PDF rendering, HTML rendering, file-copy order, or
hand-edited output changed. This separates generation logic from presentation logic.
Lexicon governance is handled as data governance. Required categories must be nonempty, optional categories remain project-owned
when declared, and the selected 40 token choice(s) stay bound to 10 configured category list(s). The slot-to-section allocation is
abstract: 3, authoring_contract: 2, configuration: 1, discussion: 3, evaluation: 5, introduction: 8, limitations: 3, methods: 6,
reproducibility: 2, results: 5, scope: 2, which lets the Methods, Results, and provenance tables state where each lexical decision
enters the manuscript.
Conditional section generation is handled before prose assembly. A disabled section does not vanish and does not borrow claims from
an enabled section; it resolves to an explicit statement that names the controlling madlib.section_conditions key. That behavior
keeps negative or excluded material visible to reviewers.
Evidence tables and visual audit figures are generated from the same config and TokenPlan.
Enabled visualizations are config-
ured_field_matrix, section_configuration_heatmap, field_origin_summary, token_injection_flow, section_token_allocation, prove-
nance_trace_map, quality_gate_matrix; they summarize field origin, token density, injection flow, section allocation, provenance,
and quality gates without adding independent claims. Figure registration is therefore a method step: every manuscript image has to
be written, registered, and validated as part of the reproducible render path.
Claim-ledger alignment is part of composition. The contribution table can make local method claims, but publication, empirical,
reader-quality, and domain-specific claims must either point to evidence or remain explicit non-claims. This is why the claim ledger,
audit rules, limitations, and authoring contract are generated beside the Methods rather than written after the fact.
Evaluation is part of the method rather than an afterthought. The config declares quality probes (Method row completeness, Field-
origin visibility, Placeholder survival, Provenance completeness, Section-switch observability, Figure registry coverage, Method-figure
alignment, Evidence cleanliness, Fork readiness, Copied-output parity, Digest invariant review, Claim-ledger alignment, Review

## Page 6

packet completeness, Fork migration sufficiency) and failure modes (Unresolved placeholder, Overclaimed generated prose, Config-
source drift, Figure provenance gap, Domain misuse, Method row drift, Field-origin opacity, Visual-method mismatch, Fork without
validators, Digest invariant drift, Claim ledger omission, Review packet incompleteness, Fork migration ambiguity); the source turns
them into tables and validation checks the rendered surface. That means methods, results, evaluation, and limitations all share one
source-owned schema instead of drifting as independent prose.
The method is organized around design principles: Configuration owns prose choices, Method surface is config-owned, Token choice is
deterministic, Field origin is evidence, Sections are conditional but visible, Visual audit follows data, Generated output is disposable,
Claim boundaries travel with prose, Forks must add validators, Invariants precede rendering, Diffs are review objects, Review packet
is a method artifact, Fork migration is part of the method. These principles prevent the Mad Lib surface from becoming a hidden
authoring channel. They require the visible manuscript to stay downstream of declared inputs, the generated outputs to remain
disposable, and the audit surface to be broad enough for a reviewer to reconstruct how a sentence reached the PDF.
The operational phases are Schema intake maps manuscript/config.yaml to MadlibConfig; Scenario declaration maps MadlibConfig to
review scenario; Field-origin inventory maps MadlibConfig and raw YAML keys to configured_field_inventory.json; Lexicon validation
maps madlib.lexicon and madlib.slots to validated slot inventory; Digest token planning maps MadlibConfig to TokenPlan; Invariant
review maps TokenPlan and method protocol to selection invariant set; Slot-to-section allocation maps TokenPlan to section token
counts; Section composition maps TokenPlan and narrative moves to manuscript variable map; Evidence table assembly maps Madli
bConfig and TokenPlan to manuscript_variables.json; Claim-ledger alignment maps MadlibConfig, generated prose, and data/clai
m_ledger.yaml to claim-aligned evidence surface; Visualization emission maps MadlibConfig, TokenPlan, and configured-field inv
entory to output/figures and figure_registry.json; Artifact emission maps MadlibConfig and TokenPlan to output/data, output/re
ports, and output/figures; Manuscript hydration maps source manuscript shells and manuscript_variables.json to hydrated Mark
down manuscript; Render maps output/manuscript to output/pdf, output/web, and output/slides; Validate and copy maps project
output directories to output/templates/template_madlib; Review packet assembly maps validated project output and copy statis
tics to review packet; Fork contract documentation maps source docs, authoring contract, and claim ledger to fork migration n
otes. Each phase has an explicit input, transformation, output, and guard. This makes the pipeline explainable at manuscript scale:
a reader can follow the path from YAML declarations to token choices, from token choices to section bodies, from section bodies to
rendered artifacts, and from rendered artifacts to validation reports.
The reviewer packet is also a method artifact. The handoff surface is hydrated Markdown, combined PDF, web output, slides, figures,
data JSON, reports, validation results, and copy statistics; a PDF alone is insufficient because it cannot show the token inventory,
field-origin inventory, figure registry, validation report, or copied-output statistics. The declared authoring obligations are Review
generated claims, Review config diffs, Extend claim evidence, Add domain validators, Rerun the full project path, Review method
invariants, Assemble reviewer packet, Write fork migration notes, which convert that packet into review work a human can actually
perform.
The claim-boundary contract is also generated. The audit-rule list contains 12 rule(s), and the contribution table binds each local
claim to a non-publication boundary. The final copy stage is a human-review handoff, not proof that the Mad Lib surface is empirically
valid or ready for a standalone release.
Fork migration closes the method. A downstream project should update config rows, source-owned composition, validators, claim-
ledger entries, and documentation before replacing exemplar vocabulary with domain claims. Without that migration work, the fork
has only changed words, not the evidential status of the generated manuscript.
3.1
Design Principles
Principle
Rationale
Manuscript effect
Configuration owns prose choices
Reviewers can inspect the declared
language surface before generation.
Large-grain manuscript variables are
generated from YAML and project source.
Method surface is config-owned
A fork should change method protocol
rows before changing generated Methods
prose.
The Methods tables and body summarize
method_protocol and pipeline_phases
from YAML.
Token choice is deterministic
A fixed seed and lexicon must produce
the same injection plan across reruns.
Token inventory rows can be regenerated
and diffed.
Field origin is evidence
A generated manuscript should
distinguish authored YAML fields from
loader defaults.
Configured-field tables and figures report
explicit and defaulted paths.
Sections are conditional but visible
Disabled material should be auditable
rather than silently absent.
Every disabled section resolves to an
explicit disabled-section body.
Visual audit follows data
Figures should explain the generated
method without becoming decorative
claims.
Cover, flow, allocation, provenance, gate,
and field-origin figures are regenerated
from artifacts.
Generated output is disposable
The durable artifact is the regeneration
contract, not hand-edited output files.
Output Markdown, PDF, HTML, and
slides are rebuilt from source inputs.

## Page 7

Principle
Rationale
Manuscript effect
Claim boundaries travel with prose
Generated text can otherwise imply
validation that no artifact supports.
Contribution, limitation, failure-mode,
and authoring tables carry boundary text.
Forks must add validators
Domain claims need domain evidence
beyond this exemplar’s generic
regeneration gates.
Scope, limitations, and authoring contract
require validators before domain claims.
Invariants precede rendering
Readers need to know which inputs are
allowed to alter generated tokens before
they inspect output.
Methods names the digest inputs and
excludes renderer state from token choice.
Diffs are review objects
Config, token inventory, manuscript
variables, figures, and copied outputs
should be diffable review surfaces.
Methods and reproducibility text describe
review packet assembly after regeneration.
Review packet is a method artifact
A PDF alone is insufficient evidence for a
generated-method exemplar.
Authoring contract and validation
sections treat data, reports, figures, and
logs as part of review.
Fork migration is part of the method
A public exemplar should teach authors
what must change before domain use.
Standalone notes, authoring obligations,
and claim-ledger boundaries name fork
responsibilities.
3.2
Operational Phases
Phase
Input
Transformation
Output
Guard
Schema intake
manuscript/config.yaml
Load paper metadata
and validate the madlib
schema before
generation.
MadlibConfig
config parser tests
Scenario declaration
MadlibConfig
Summarize local scope,
enabled sections, claim
boundaries, and review
handoff expectations.
review scenario
method protocol and
contribution table tests
Field-origin inventory
MadlibConfig and raw Y
AML keys
Classify supported
paths as explicit or
defaulted.
configured_field_inven
tory.json
configured-field
inventory tests
Lexicon validation
madlib.lexicon and mad
lib.slots
Reject empty required
categories and slot
references to missing
categories.
validated slot invento
ry
malformed-config tests
Digest token planning
MadlibConfig
Hash seed, slot name,
category, ordinal, and
category inventory.
TokenPlan
seed-stability tests
Invariant review
TokenPlan and method p
rotocol
Confirm allowed
token-choice inputs are
documented and
isolated from renderer
state.
selection invariant se
t
token determinism and
Methods prose tests
Slot-to-section
allocation
TokenPlan
Assign each selected
token to its configured
manuscript section.
section token counts
provenance and
allocation tests
Section composition
TokenPlan and narrativ
e moves
Build conditional
section bodies, titles,
and evidence tables.
manuscript variable ma
p
placeholder-coverage
tests
Evidence table assembly
MadlibConfig and Token
Plan
Render protocol, phase,
audit, token, section,
provenance, and
configured-field tables.
manuscript_variables.j
son
composition table tests
Claim-ledger alignment
MadlibConfig, generate
d prose, and data/clai
m_ledger.yaml
Check local claims and
non-claims against
source-owned evidence
rows.
claim-aligned evidence
surface
claim ledger and
evidence registry review

## Page 8

Phase
Input
Transformation
Output
Guard
Visualization emission
MadlibConfig, TokenPla
n, and configured-fiel
d inventory
Generate cover, flow,
allocation, provenance,
gate, and
configured-field figures.
output/figures and fig
ure_registry.json
nonblank figure tests
Artifact emission
MadlibConfig and Token
Plan
Write inventory, section
plan, injection trace,
summary, cover
overview, manuscript
figures, and registry.
output/data, output/re
ports, and output/figu
res
artifact-writing tests
Manuscript hydration
source manuscript shel
ls and manuscript_vari
ables.json
Resolve large-grain
placeholders into
output/manuscript.
hydrated Markdown manu
script
unresolved-token scan
Render
output/manuscript
Render PDF, HTML,
and slides through the
shared template
pipeline.
output/pdf, output/web
, and output/slides
render command
Validate and copy
project output directo
ries
Validate files, registries,
evidence, overlays, and
copied deliverables.
output/templates/templ
ate_madlib
validation and copy
commands
Review packet assembly
validated project outp
ut and copy statistics
Group manuscript, web,
slides, figures, data,
reports, validation
results, and copy
statistics as the review
surface.
review packet
copied-output
validation
Fork contract
documentation
source docs, authoring
contract, and claim le
dger
State which config,
source, test, validator,
and evidence surfaces a
fork must change before
domain claims.
fork migration notes
documentation and
claim-ledger tests
3.3
Protocol Steps
Step
Action
Evidence
Output
Ingest declared manuscript
schema
Parse paper metadata and the
madlib block from
manuscript/config.yaml before
any prose or figures are
composed.
Config validation tests and
MadlibConfig construction
from the committed YAML.
MadlibConfig
Declare review scenario
Name the manuscript scope,
local claim boundary, enabled
sections, and intended
reviewer handoff before token
generation.
section_plan.json,
contribution table, and
authoring contract rows.
review scenario
Track field origin
Record every supported
madlib path as explicit when it
appears in YAML or defaulted
when the loader supplies it.
configured_field_inventory.json
and configured-field origin
tests.
explicit/default path invent
ory
Govern lexicon categories
Reject empty required
categories, preserve
project-owned optional
categories, and treat every
lexical list as source data.
Malformed-config tests and
lexicon rows in
token_inventory.json.
validated lexicon
Construct digest selection
material
Combine seed, slot name,
category, ordinal, and the full
category inventory into a
deterministic digest input.
Seed-stability and
category-sensitivity tests.
digest input records

## Page 9

Step
Action
Evidence
Output
Record selection invariants
State the invariant inputs that
are allowed to change token
choices and the renderer state
that is not allowed to affect
them.
Token determinism tests and
Methods digest prose.
selection invariant set
Expand slot declarations
Resolve every slot count into
one or more token choices and
assign each selected value to
its configured manuscript
section.
TokenPlan construction,
provenance trace, and section
allocation figure.
TokenPlan
Apply section conditions
Evaluate section switches
before prose assembly so
disabled sections receive
explicit disabled-section
bodies.
Section-condition tests and
out-
put/data/section_plan.json.
enabled section set
Compose conditional IMRAD
bodies
Build large-grain section
variables from narrative moves,
selected tokens, section
switches, and local claim
boundaries.
Generated manuscript
variables and hydrated
output/manuscript Markdown.
section variables
Assemble evidence tables
Render method protocol,
pipeline phase, design
principle, audit rule, token
inventory, section plan, and
provenance tables from config
and TokenPlan.
Composition tests and
generated
manuscript_variables.json
table entries.
Markdown evidence tables
Align claims with evidence
ledger
Keep generated method,
visualization, and
publication-boundary claims
tied to config, source modules,
generated artifacts, or explicit
non-claim boundaries.
data/claim_ledger.yaml and
evidence registry validation.
claim-aligned evidence surfa
ce
Generate visual audit surface
Write the cover overview,
token-injection flow, section
allocation, provenance trace,
quality gate, and
configured-field figures from
generated data.
Nonblank figure tests and
../figures/figure_registry.json.
registered figure set
Emit machine-readable
artifacts
Write token inventory, section
plan, configured-field
inventory, injection trace,
summary reports, validation
inputs, and figure registry.
Artifact-writing tests, artifact
manifest, and validation
reports.
output/data, output/reports,
and output/figures
Hydrate manuscript shell
Write
manuscript_variables.json and
resolve the source Markdown
shells into output/manuscript.
Unresolved-token scan and
render validation.
output/manuscript
Render and validate
deliverables
Render PDF, HTML, and
slides through shared
infrastructure, then validate
PDFs, Markdown, figure
registry, evidence registry,
design overlays, and artifact
manifest.
Stage 03 render log and Stage
04 validation report.
validated project output
Assemble reviewer packet
Treat hydrated Markdown,
rendered PDF, web output,
slides, figures, data, reports,
validation logs, and copied
output statistics as one review
packet.
Stage 04 validation report and
Stage 05
output_statistics.json.
review packet

## Page 10

Step
Action
Evidence
Output
Copy review surface
Copy validated deliverables
into out-
put/templates/template_madlib
only after validation passes.
Stage 05 copy statistics and
copied-output validation.
copied publication-review bu
ndle
Document fork migration
Record what downstream
authors must change when
moving from exemplar token
injection to a domain-specific
report.
README, STANDALONE
notes, manuscript README,
Authoring Contract, and claim
ledger boundary rows.
fork migration notes
3.4
Section Plan
Section
Render title
Enabled
Token choices
Narrative moves
abstract
Abstract
True
3
state the manuscript-
generation problem,
name the
deterministic
intervention,
summarize the audit
surface
introduction
Introduction:
Lexicon as Data and
Manuscript as Build
Artifact
True
8
separate playful Mad
Lib syntax from
research claims,
identify drift
between prose and
source data, frame
configuration as an
inspectable dataset,
position conditional
prose as a
reproducibility
problem

## Page 11

Section
Render title
Enabled
Token choices
Narrative moves
methods
Methods:
Source-Owned Token
Injection and
Conditional IMRAD
Assembly
True
6
validate config before
composition, declare
review scenario and
method invariants,
separate explicit
YAML fields from
loader defaults,
govern lexicon
categories as
reviewable data,
expand slots through
seeded digest
selection, allocate
slot choices to
enabled manuscript
sections, compose
conditional section
bodies, assemble
method evidence
tables and visual
audit figures, align
generated claims
with the claim ledger,
emit evidence
artifacts before
rendering, verify that
no unresolved
placeholders remain,
assemble a reviewer
packet from
regenerated artifacts,
preserve human
review before
publication claims
results
Results: Provenance,
Density, and
Resolved Manuscript
Surface
True
5
report token density,
show resolved section
coverage, bind every
manuscript token to
provenance, connect
the figure and
inventory to the
same plan
discussion
Discussion:
Accountability
Boundaries for
Generated Prose
True
3
bound the scholarly
claim, describe useful
adaptation cases,
name misuse modes,
preserve human
authorship
responsibility
configuration
Configuration:
Schema-Controlled
Lexicon, Slots, and
Narrative Moves
True
1
document schema
ownership, show title
and switch behavior,
record generated
counts from code
evaluation
Evaluation: Gate
Criteria, QA Probes,
and Failure
Discovery
True
5
name readiness
criteria, connect
criteria to artifacts,
separate local checks
from publication
readiness, make
failure probes visible

## Page 12

Section
Render title
Enabled
Token choices
Narrative moves
reproducibility
Reproducibility:
Seeded Regeneration
and Artifact Trace
True
2
fix seed and config
hash, write
machine-readable
artifacts, rerender
through the shared
pipeline, copy
outputs only after
validation
limitations
Limitations:
Non-Claims, Misuse
Modes, and Human
Review
True
3
state non-claims,
identify misuse
modes, preserve
human review,
require domain
validators for domain
claims
scope
Scope: Related
Generators and
Responsible Forking
True
2
distinguish
generation from
truth, limit
publication claims,
point to local
evidence, explain
responsible forking
authoring_contract
Authoring Contract:
Human Review and
Forking Obligations
True
2
state human
responsibilities,
name fork
obligations, connect
review to generated
evidence, require
domain validators
before domain
claims, document
fork migration notes
3.5
Audit Rules
Rule
Enforcement surface
R1
Every manuscript placeholder must be generated by source code.
R2
Every generated token choice must carry category and
config-key provenance.
R3
Every method protocol row must identify an action, evidence
artifact, and output.
R4
Every pipeline phase must identify an input, transformation,
output, and guard.
R5
Every visible configured field must be classified as explicit or
defaulted.
R6
Every disabled section must resolve to an explicit
disabled-section body.
R7
Every figure reference must be backed by a generated figure
registry entry.
R8
Every fork that adds domain claims must add domain validators
and claim-ledger evidence.
R9
Every publication claim must stay local unless a live DOI or
release exists.
R10
Every token-selection explanation must name only seed, slot,
category, ordinal, and ordered category inventory as digest
inputs.
R11
Every review handoff must include generated data, reports,
figures, validation results, and copy statistics alongside PDF or
HTML.

## Page 13

Rule
Enforcement surface
R12
Every fork migration note must name config, source, test,
validator, pipeline, and claim-ledger obligations.

## Page 14

Figure 1: Deterministic token-injection flow
4
Results: Provenance, Density, and Resolved Manuscript Surface
The generated plan enables 11 of 11 manuscript sections and fills 40 token choice(s). Category density is adjectives: 2, artifacts: 8,
audiences: 4, constraints: 3, failures: 3, measures: 6, methods: 1, nouns: 5, qualities: 3, verbs: 5. The result figures are generated
from the same token plan that writes the inventory table, so visual and tabular claims share one source.
The configured results moves are report token density, show resolved section coverage, bind every manuscript token to provenance,
and connect the figure and inventory to the same plan. The important result is therefore not a surprising word choice; it is the
survival of traceability through a complete render path. Each token row records the variable, category, selected value, section, and
config pointer that produced it.
The resolved manuscript also demonstrates the intended failure boundary. If a manuscript placeholder is added without a correspond-
ing variable, the project test suite detects it. If a figure is referenced without registry support, the output validator reports it. If a
generated number lacks evidence support, the evidence registry gate reports it before the copied output stage packages deliverables.
Visualization is enabled for configured_field_matrix, section_configuration_heatmap, field_origin_summary, token_injection_flow,
section_token_allocation, provenance_trace_map, quality_gate_matrix. The configured-field figures are generated from the same
explicit/default path inventory written to JSON; the pipeline, allocation, provenance, and gate figures are generated from the same
token plan and QA schema.
4.1
Token Inventory
Variable
Category
Value
Section
Source
STUDY_ADJECTIVE
adjectives
reviewable
abstract
manuscript/config.yaml#madl
STUDY_NOUN
nouns
pipeline
abstract
manuscript/config.yaml#madl
STUDY_VERB
verbs
hydrate
abstract
manuscript/config.yaml#madl
INTRO_NOUNS_1
nouns
protocol
introduction
manuscript/config.yaml#madl
INTRO_NOUNS_2
nouns
section
introduction
manuscript/config.yaml#madl
INTRO_NOUNS_3
nouns
lexicon
introduction
manuscript/config.yaml#madl
INTRO_NOUNS_4
nouns
artifact
introduction
manuscript/config.yaml#madl
INTRO_VERBS_1
verbs
condition
introduction
manuscript/config.yaml#madl
INTRO_VERBS_2
verbs
bind
introduction
manuscript/config.yaml#madl
INTRO_VERBS_3
verbs
bind
introduction
manuscript/config.yaml#madl
INTRO_VERBS_4
verbs
compose
introduction
manuscript/config.yaml#madl
METHOD_NAME
methods
conditional section
hydration
methods
manuscript/config.yaml#madl
METHOD_CONSTRAINT
constraints
publication claims stay
local until release
methods
manuscript/config.yaml#madl

## Page 15

Variable
Category
Value
Section
Source
METHOD_ARTIFACT_1
artifacts
token-injection flow
methods
manuscript/config.yaml#madl
METHOD_ARTIFACT_2
artifacts
quality-gate matrix
methods
manuscript/config.yaml#madl
METHOD_QUALITY_1
qualities
claim humility
methods
manuscript/config.yaml#madl
METHOD_QUALITY_2
qualities
render readiness
methods
manuscript/config.yaml#madl
RESULT_MEASURE_1
measures
provenance coverage
results
manuscript/config.yaml#madl
RESULT_MEASURE_2
measures
evidence registry
cleanliness
results
manuscript/config.yaml#madl
RESULT_MEASURE_3
measures
category density
results
manuscript/config.yaml#madl
RESULT_ARTIFACT_1
artifacts
configured-field figures
results
manuscript/config.yaml#madl
RESULT_ARTIFACT_2
artifacts
token inventory
results
manuscript/config.yaml#madl
DISCUSSION_ADJECTIVE
adjectives
auditable
discussion
manuscript/config.yaml#madl
DISCUSSION_AUDIENCE_1
audiences
pipeline maintainers
discussion
manuscript/config.yaml#madl
DISCUSSION_AUDIENCE_2
audiences
research educators
discussion
manuscript/config.yaml#madl
CONFIG_CONSTRAINT
constraints
disabled sections retain
explicit traceability
configuration
manuscript/config.yaml#madl
EVALUATION_MEASURE_1
measures
copied output readiness
evaluation
manuscript/config.yaml#madl
EVALUATION_MEASURE_2
measures
figure registry
completeness
evaluation
manuscript/config.yaml#madl
EVALUATION_MEASURE_3
measures
category density
evaluation
manuscript/config.yaml#madl
EVALUATION_ARTIFACT_1
artifacts
manuscript variable
map
evaluation
manuscript/config.yaml#madl
EVALUATION_ARTIFACT_2
artifacts
provenance trace map
evaluation
manuscript/config.yaml#madl
REPRODUCIBILITY_ARTIFA
CT_1
artifacts
section plan
reproducibility
manuscript/config.yaml#madl
REPRODUCIBILITY_ARTIFA
CT_2
artifacts
manuscript variable
map
reproducibility
manuscript/config.yaml#madl
LIMITATION_FAILURE_1
failures
domain misuse
limitations
manuscript/config.yaml#madl
LIMITATION_FAILURE_2
failures
overclaimed generated
prose
limitations
manuscript/config.yaml#madl
LIMITATION_FAILURE_3
failures
figure provenance gap
limitations
manuscript/config.yaml#madl
SCOPE_CONSTRAINT
constraints
all lexicon entries live in
config
scope
manuscript/config.yaml#madl
SCOPE_AUDIENCE
audiences
pipeline maintainers
scope
manuscript/config.yaml#madl
AUTHORING_AUDIENCE
audiences
manuscript reviewers
authoring_contract
manuscript/config.yaml#madl
AUTHORING_QUALITY
qualities
render readiness
authoring_contract
manuscript/config.yaml#madl
4.2
Provenance Matrix
Section
Token variables
Source categories
Abstract
STUDY_ADJECTIVE, STUDY_NOUN, STUDY_VERB
adjectives, nouns, verbs
Introduction: Lexicon as Data and
Manuscript as Build Artifact
INTRO_NOUNS_1, INTRO_NOUNS_2, INTRO_NOUNS_3,
INTRO_NOUNS_4, INTRO_VERBS_1, INTRO_VERBS_2,
INTRO_VERBS_3, INTRO_VERBS_4
nouns, verbs
Methods: Source-Owned Token
Injection and Conditional IMRAD
Assembly
METHOD_NAME, METHOD_CONSTRAINT, METHOD_ARTIFACT_
1, METHOD_ARTIFACT_2, METHOD_QUALITY_1, METHOD_QU
ALITY_2
artifacts, constraints, methods,
qualities
Results: Provenance, Density, and
Resolved Manuscript Surface
RESULT_MEASURE_1, RESULT_MEASURE_2, RESULT_MEASU
RE_3, RESULT_ARTIFACT_1, RESULT_ARTIFACT_2
artifacts, measures
Discussion: Accountability
Boundaries for Generated Prose
DISCUSSION_ADJECTIVE, DISCUSSION_AUDIENCE_1, DIS
CUSSION_AUDIENCE_2
adjectives, audiences
Configuration: Schema-Controlled
Lexicon, Slots, and Narrative Moves
CONFIG_CONSTRAINT
constraints
Evaluation: Gate Criteria, QA
Probes, and Failure Discovery
EVALUATION_MEASURE_1, EVALUATION_MEASURE_2, EVAL
UATION_MEASURE_3, EVALUATION_ARTIFACT_1, EVALUAT
ION_ARTIFACT_2
artifacts, measures
Reproducibility: Seeded Regeneration
and Artifact Trace
REPRODUCIBILITY_ARTIFACT_1, REPRODUCIBILITY_ART
IFACT_2
artifacts
Limitations: Non-Claims, Misuse
Modes, and Human Review
LIMITATION_FAILURE_1, LIMITATION_FAILURE_2, LIMI
TATION_FAILURE_3
failures

## Page 16

Section
Token variables
Source categories
Scope: Related Generators and
Responsible Forking
SCOPE_CONSTRAINT, SCOPE_AUDIENCE
audiences, constraints
Authoring Contract: Human Review
and Forking Obligations
AUTHORING_AUDIENCE, AUTHORING_QUALITY
audiences, qualities

## Page 17

Figure 2: Token category density
Figure 3: Section token allocation

## Page 18

Figure 4: Provenance trace map

## Page 19

5
Discussion: Accountability Boundaries for Generated Prose
The reviewable result is intentionally modest. The exemplar does not claim that random lexical replacement creates scholarship,
discovers facts, or substitutes for authorship. It shows that a conditional text generator can be made accountable to configuration,
tests, and render-time validation.
The configured discussion moves are bound the scholarly claim, describe useful adaptation cases, name misuse modes, and preserve
human authorship responsibility. Useful adaptations include templated empirical reports, structured review memos, classroom ex-
ercises, and manuscript sections that need to toggle across protocols. Risky adaptations include hiding weak claims behind fluent
generated phrasing or allowing a token category to imply evidence that the project never produced.
The pattern scales only when authors preserve the same ownership boundary. Lexicons should be small enough to review, categories
should be named for their manuscript function, and section switches should state what has been removed. When richer language is
needed, the next layer should add domain-specific validators rather than relaxing provenance.

## Page 20

6
Configuration: Schema-Controlled Lexicon, Slots, and Narrative Moves
The active composition depth is deep. The lexicon exposes 10 category list(s), and the slot declaration expands to 40 concrete token
choice(s). Section switches are evaluated before composition so disabled sections cannot silently borrow enabled-section claims.
Configuration owns more than vocabulary. It also owns section titles, narrative moves, method protocol rows, contribution claims, and
audit rules. That makes the manuscript shape visible in one YAML file while preserving the rule that source code, not hand-edited
output, performs the composition.
The tables in this section expose the declared surface that controls rendering. They are useful during review because a title change,
slot expansion, or disabled section appears as a small config diff and a regenerated artifact diff rather than as scattered prose edits.
The configured-field inventory separates 125 explicit YAML path(s) from 11 loader-defaulted path(s). That distinction matters for
forks: a field that appears in the rendered manuscript may be intentionally authored in config.yaml, or it may be a documented
default inherited from the template.
Figure 5: Configured field origin matrix
6.1
Declared Section Titles
Section key
Rendered title
Enabled
abstract
Abstract
True
introduction
Introduction: Lexicon as Data and
Manuscript as Build Artifact
True
methods
Methods: Source-Owned Token
Injection and Conditional IMRAD
Assembly
True
results
Results: Provenance, Density, and
Resolved Manuscript Surface
True
discussion
Discussion: Accountability
Boundaries for Generated Prose
True
configuration
Configuration: Schema-Controlled
Lexicon, Slots, and Narrative Moves
True
evaluation
Evaluation: Gate Criteria, QA
Probes, and Failure Discovery
True
reproducibility
Reproducibility: Seeded Regeneration
and Artifact Trace
True
limitations
Limitations: Non-Claims, Misuse
Modes, and Human Review
True

## Page 21

Section key
Rendered title
Enabled
scope
Scope: Related Generators and
Responsible Forking
True
authoring_contract
Authoring Contract: Human Review
and Forking Obligations
True
6.2
Configuration Counts
• Seed: 431
• Composition depth: deep
• Lexicon categories: 10
• Slot rules: 22
• Token choices: 40
• Enabled sections: 11
• Method steps: 18
• Design principles: 13
• Pipeline phases: 17
• Quality probes: 14
• Authoring obligations: 8
• Explicit configured paths: 125
• Defaulted configured paths: 11
• Enabled visualization flags: 7
• Section-level configured paths: 33
• Lexicon-level configured paths: 10
• Slot-level configured paths: 66
• Narrative moves: 52
• Audit rules: 12
• Contribution claims: 9
6.3
Configured Field Summary
Measure
Count
Total tracked field paths
136
Explicit YAML paths
125
Loader-defaulted paths
11
Enabled visualization flags
7
Section-level paths
33
Lexicon-level paths
10
Slot-level paths
66
Visualization-control paths
9
Top-level schema paths
18
6.4
Configured Field Inventory
Path
Origin
Scope
Summary
madlib
explicit
schema
configured field
madlib.audit_rules
explicit
schema
12 entries
madlib.authoring_obligations
explicit
schema
8 entries
madlib.composition_depth
explicit
schema
deep
madlib.contribution_claims
explicit
schema
9 entries
madlib.design_principles
explicit
schema
13 entries
madlib.evaluation_criteria
explicit
schema
7 entries
madlib.failure_modes
explicit
schema
13 entries

## Page 22

Path
Origin
Scope
Summary
madlib.hypothesis
explicit
schema
Deterministic lexical injection
can generate a complete
conditional IMRAD
manuscript while preserving
token provenance, section
intent, and audit-ready
method evidence.
madlib.lexicon
explicit
schema
10 categories
madlib.lexicon.adjectives
explicit
lexicon
7 tokens
madlib.lexicon.artifacts
explicit
lexicon
12 tokens
madlib.lexicon.audiences
explicit
lexicon
4 tokens
madlib.lexicon.constraints
explicit
lexicon
5 tokens
madlib.lexicon.failures
explicit
lexicon
5 tokens
madlib.lexicon.measures
explicit
lexicon
8 tokens
madlib.lexicon.methods
explicit
lexicon
5 tokens
madlib.lexicon.nouns
explicit
lexicon
7 tokens
madlib.lexicon.qualities
explicit
lexicon
5 tokens
madlib.lexicon.verbs
explicit
lexicon
7 tokens
madlib.method_protocol
explicit
schema
18 entries
madlib.narrative_moves
explicit
schema
configured field
madlib.narrative_moves.abstr
act
explicit
section
3 moves
madlib.narrative_moves.autho
ring_contract
explicit
section
5 moves
madlib.narrative_moves.confi
guration
explicit
section
3 moves
madlib.narrative_moves.discu
ssion
explicit
section
4 moves
madlib.narrative_moves.evalu
ation
explicit
section
4 moves
madlib.narrative_moves.intro
duction
explicit
section
4 moves
madlib.narrative_moves.limit
ations
explicit
section
4 moves
madlib.narrative_moves.metho
ds
explicit
section
13 moves
madlib.narrative_moves.repro
ducibility
explicit
section
4 moves
madlib.narrative_moves.resul
ts
explicit
section
4 moves
madlib.narrative_moves.scope
explicit
section
4 moves
madlib.pipeline_phases
explicit
schema
17 entries
madlib.quality_probes
explicit
schema
14 entries
madlib.section_conditions
explicit
schema
configured field
madlib.section_conditions.ab
stract
explicit
section
enabled
madlib.section_conditions.au
thoring_contract
explicit
section
enabled
madlib.section_conditions.co
nfiguration
explicit
section
enabled
madlib.section_conditions.di
scussion
explicit
section
enabled
madlib.section_conditions.ev
aluation
explicit
section
enabled
madlib.section_conditions.in
troduction
explicit
section
enabled
madlib.section_conditions.li
mitations
explicit
section
enabled
madlib.section_conditions.me
thods
explicit
section
enabled

## Page 23

Path
Origin
Scope
Summary
madlib.section_conditions.re
producibility
explicit
section
enabled
madlib.section_conditions.re
sults
explicit
section
enabled
madlib.section_conditions.sc
ope
explicit
section
enabled
madlib.section_titles
explicit
schema
configured field
madlib.section_titles.abstra
ct
explicit
section
Abstract
madlib.section_titles.author
ing_contract
explicit
section
Authoring Contract: Human
Review and Forking
Obligations
madlib.section_titles.config
uration
explicit
section
Configuration:
Schema-Controlled Lexicon,
Slots, and Narrative Moves
madlib.section_titles.discus
sion
explicit
section
Discussion: Accountability
Boundaries for Generated
Prose
madlib.section_titles.evalua
tion
explicit
section
Evaluation: Gate Criteria, QA
Probes, and Failure Discovery
madlib.section_titles.introd
uction
explicit
section
Introduction: Lexicon as Data
and Manuscript as Build
Artifact
madlib.section_titles.limita
tions
explicit
section
Limitations: Non-Claims,
Misuse Modes, and Human
Review
madlib.section_titles.method
s
explicit
section
Methods: Source-Owned
Token Injection and
Conditional IMRAD Assembly
madlib.section_titles.reprod
ucibility
explicit
section
Reproducibility: Seeded
Regeneration and Artifact
Trace
madlib.section_titles.result
s
explicit
section
Results: Provenance, Density,
and Resolved Manuscript
Surface
madlib.section_titles.scope
explicit
section
Scope: Related Generators
and Responsible Forking
madlib.seed
explicit
schema
431
madlib.slots
explicit
schema
22 slot rules, 40 token choices
madlib.slots.authoring_audie
nce
explicit
slot
audiences ->
authoring_contract (1)
madlib.slots.authoring_audie
nce.count
defaulted
slot
1
madlib.slots.authoring_audie
nce.section
explicit
slot
authoring_contract
madlib.slots.authoring_quali
ty
explicit
slot
qualities ->
authoring_contract (1)
madlib.slots.authoring_quali
ty.count
defaulted
slot
1
madlib.slots.authoring_quali
ty.section
explicit
slot
authoring_contract
madlib.slots.config_constrai
nt
explicit
slot
constraints -> configuration
(1)
madlib.slots.config_constrai
nt.count
defaulted
slot
1
madlib.slots.config_constrai
nt.section
explicit
slot
configuration
madlib.slots.discussion_adje
ctive
explicit
slot
adjectives -> discussion (1)
madlib.slots.discussion_adje
ctive.count
defaulted
slot
1

## Page 24

Path
Origin
Scope
Summary
madlib.slots.discussion_adje
ctive.section
explicit
slot
discussion
madlib.slots.discussion_audi
ence
explicit
slot
audiences -> discussion (2)
madlib.slots.discussion_audi
ence.count
explicit
slot
2
madlib.slots.discussion_audi
ence.section
explicit
slot
discussion
madlib.slots.evaluation_arti
fact
explicit
slot
artifacts -> evaluation (2)
madlib.slots.evaluation_arti
fact.count
explicit
slot
2
madlib.slots.evaluation_arti
fact.section
explicit
slot
evaluation
madlib.slots.evaluation_meas
ure
explicit
slot
measures -> evaluation (3)
madlib.slots.evaluation_meas
ure.count
explicit
slot
3
madlib.slots.evaluation_meas
ure.section
explicit
slot
evaluation
madlib.slots.intro_nouns
explicit
slot
nouns -> introduction (4)
madlib.slots.intro_nouns.cou
nt
explicit
slot
4
madlib.slots.intro_nouns.sec
tion
explicit
slot
introduction
madlib.slots.intro_verbs
explicit
slot
verbs -> introduction (4)
madlib.slots.intro_verbs.cou
nt
explicit
slot
4
madlib.slots.intro_verbs.sec
tion
explicit
slot
introduction
madlib.slots.limitation_fail
ure
explicit
slot
failures -> limitations (3)
madlib.slots.limitation_fail
ure.count
explicit
slot
3
madlib.slots.limitation_fail
ure.section
explicit
slot
limitations
madlib.slots.method_artifact
explicit
slot
artifacts -> methods (2)
madlib.slots.method_artifact
.count
explicit
slot
2
madlib.slots.method_artifact
.section
explicit
slot
methods
madlib.slots.method_constrai
nt
explicit
slot
constraints -> methods (1)
madlib.slots.method_constrai
nt.count
defaulted
slot
1
madlib.slots.method_constrai
nt.section
explicit
slot
methods
madlib.slots.method_name
explicit
slot
methods -> methods (1)
madlib.slots.method_name.cou
nt
defaulted
slot
1
madlib.slots.method_name.sec
tion
explicit
slot
methods
madlib.slots.method_quality
explicit
slot
qualities -> methods (2)
madlib.slots.method_quality.
count
explicit
slot
2
madlib.slots.method_quality.
section
explicit
slot
methods
madlib.slots.reproducibility
_artifact
explicit
slot
artifacts -> reproducibility (2)
madlib.slots.reproducibility
_artifact.count
explicit
slot
2

## Page 25

Path
Origin
Scope
Summary
madlib.slots.reproducibility
_artifact.section
explicit
slot
reproducibility
madlib.slots.result_artifact
explicit
slot
artifacts -> results (2)
madlib.slots.result_artifact
.count
explicit
slot
2
madlib.slots.result_artifact
.section
explicit
slot
results
madlib.slots.result_measure
explicit
slot
measures -> results (3)
madlib.slots.result_measure.
count
explicit
slot
3
madlib.slots.result_measure.
section
explicit
slot
results
madlib.slots.scope_audience
explicit
slot
audiences -> scope (1)
madlib.slots.scope_audience.
count
defaulted
slot
1
madlib.slots.scope_audience.
section
explicit
slot
scope
madlib.slots.scope_constrain
t
explicit
slot
constraints -> scope (1)
madlib.slots.scope_constrain
t.count
defaulted
slot
1
madlib.slots.scope_constrain
t.section
explicit
slot
scope
madlib.slots.study_adjective
explicit
slot
adjectives -> abstract (1)
madlib.slots.study_adjective
.count
defaulted
slot
1
madlib.slots.study_adjective
.section
explicit
slot
abstract
madlib.slots.study_noun
explicit
slot
nouns -> abstract (1)
madlib.slots.study_noun.coun
t
defaulted
slot
1
madlib.slots.study_noun.sect
ion
explicit
slot
abstract
madlib.slots.study_verb
explicit
slot
verbs -> abstract (1)
madlib.slots.study_verb.coun
t
defaulted
slot
1
madlib.slots.study_verb.sect
ion
explicit
slot
abstract
madlib.visualizations
explicit
visualization
enabled
madlib.visualizations.config
ured_field_matrix
explicit
visualization
true
madlib.visualizations.enable
d
explicit
visualization
true
madlib.visualizations.field_
origin_summary
explicit
visualization
true
madlib.visualizations.proven
ance_trace_map
explicit
visualization
true
madlib.visualizations.qualit
y_gate_matrix
explicit
visualization
true
madlib.visualizations.sectio
n_configuration_heatmap
explicit
visualization
true
madlib.visualizations.sectio
n_token_allocation
explicit
visualization
true
madlib.visualizations.token_
injection_flow
explicit
visualization
true

## Page 26

Figure 6: Section configuration heatmap
Figure 7: Field origin summary

## Page 27

7
Evaluation: Gate Criteria, QA Probes, and Failure Discovery
The evaluation section is configured to name readiness criteria, connect criteria to artifacts, separate local checks from publication
readiness, and make failure probes visible. The local readiness surface is not a human preference score; it is a set of deterministic
checks that connect generated manuscript claims to source files and pipeline gates.
The active criteria use analysis, copy, pytest, validation as gate labels and inspect artifacts such as token-injection flow, quality-gate
matrix, configured-field figures. A passing run means the exemplar is locally render-ready: placeholders resolve, token provenance is
present, figure references are registered, evidence scanning has not found unsupported numbers, and project design overlays remain
internally consistent.
That readiness is deliberately narrower than publication readiness. A local pass does not imply a standalone DOI, external release,
reader preference result, or empirical validation. It means the tracked project tree can regenerate the committed artifact surface
through its declared pipeline.
The QA probes are Method row completeness, Field-origin visibility, Placeholder survival, Provenance completeness, Section-switch
observability, Figure registry coverage, Method-figure alignment, Evidence cleanliness, Fork readiness, Copied-output parity, Digest
invariant review, Claim-ledger alignment, Review packet completeness, Fork migration sufficiency. They are phrased as questions so
they can be reused by reviewers and by forks of the exemplar: did the placeholder disappear, did the provenance survive, did the
figure registry cover every reference, and did copied outputs preserve the same evidence surface that validation inspected?
Figure 8: Quality gate matrix
7.1
Evaluation Criteria
Criterion
Target
Evidence
Gate
Placeholder resolution
No unresolved uppercase
manuscript placeholders
remain in output/manuscript
or rendered web output.
tests/test_manuscript_variables.py
and rg unresolved-token scan.
pytest
Token provenance coverage
Every selected token maps to
category, selected value,
section, and config pointer.
output/reports/injection_trace.json
and out-
put/data/token_inventory.json.
analysis

## Page 28

Criterion
Target
Evidence
Gate
Figure registry integrity
Every referenced figure label is
present in
../figures/figure_registry.json.
scripts/04_validate_output.py
figure registry check.
validation
Evidence registry cleanliness
Generated manuscript
numbers and claims pass the
project evidence registry.
output/reports/evidence_registry.json.
validation
Copied-output readiness
PDF, HTML, slides, figures,
data, and reports copy into out-
put/templates/template_madlib.
scripts/05_copy_outputs.py
output statistics.
copy
Reviewer packet completeness
Hydrated Markdown, rendered
PDF, web output, slides,
figures, data, reports,
validation results, and copy
statistics are all present for
review.
Stage 04 validation report and
Stage 05
output_statistics.json.
copy
Method-invariant traceability
Token choices can be
explained only by seed, slot,
category, ordinal, and ordered
category inventory.
tests/test_tokens.py and
generated Methods digest
prose.
pytest
7.2
Quality Probes
Probe
Question
Passing signal
Artifact
Method row completeness
Does the protocol table cover
schema intake, token planning,
composition, figures,
validation, copy, and review
handoff?
method_protocol includes
rows for every major pipeline
responsibility.
manuscript/config.yaml and o
utput/data/section_plan.json
Field-origin visibility
Can a reviewer tell which
visible fields were authored
and which were defaulted?
Configured-field inventory and
summary tables report explicit
and defaulted paths.
output/data/configured_field
_inventory.json
Placeholder survival
Did any source token survive
hydration?
No uppercase placeholders are
found in generated manuscript
or web files.
output/manuscript and output
/web
Provenance completeness
Can every selected token be
traced to a category, section,
value, and config key?
The injection trace and token
inventory contain one row for
each generated token.
output/reports/injection_tra
ce.json
Section-switch observability
Does a disabled section resolve
to a visible explanation?
Disabled section bodies cite
their controlling section
condition.
output/data/section_plan.jso
n
Figure registry coverage
Does every manuscript figure
reference have a generated
registry entry?
Figure registry validation
passes.
../figures/figure_registry.j
son
Method-figure alignment
Do method figures describe
generated data rather than
decorative or unsupported
claims?
Figure registry captions,
nonblank PNG tests, and
manual visual QA align with
artifact data.
output/figures
Evidence cleanliness
Do generated claims stay
within local evidence
boundaries?
Evidence registry validation
passes without unsupported
claims.
output/reports/evidence_regi
stry.json
Fork readiness
Does the authoring contract
tell downstream forks what to
extend before adding domain
claims?
Authoring obligations cite
config diffs, claim ledger
updates, validators, and full
reruns.
output/manuscript/10_authori
ng_contract.md
Copied-output parity
Did copied deliverables
preserve the validated project
output surface?
Copy-stage statistics include
PDF, HTML, slides, figures,
data, and reports.
output/templates/template_ma
dlib

## Page 29

Probe
Question
Passing signal
Artifact
Digest invariant review
Are the allowed token-selection
inputs documented and
protected by tests?
Methods prose names the
digest inputs and token tests
prove seed/category
sensitivity.
src/tokens.py and output/man
uscript/02_methodology.md
Claim-ledger alignment
Do method and
documentation claims point to
config, source, generated
artifacts, or explicit non-claim
boundaries?
Claim-ledger rows cover
expanded method protocol
and fork-validator boundaries.
data/claim_ledger.yaml
Review packet completeness
Can a reviewer inspect every
output surface needed to audit
the method?
Copied outputs include
manuscript, web, slides,
figures, data, reports,
validation, and copy statistics.
output/templates/template_ma
dlib and output/reports/outp
ut_statistics.json
Fork migration sufficiency
Does the documentation tell
forks which surfaces to change
before adding domain claims?
README, STANDALONE,
manuscript README, and
Authoring Contract list config,
source, test, validator, and
claim-ledger obligations.
README.md, STANDALONE.md, ma
nuscript/README.md, and outp
ut/manuscript/10_authoring_c
ontract.md

## Page 30

8
Reproducibility: Seeded Regeneration and Artifact Trace
Re-running generation with seed 431 and the same lexicon produces the same token plan.
The artifact set records toke
n_inventory.json,
section_plan.json,
injection_trace.json,
manuscript_variables.json,
figure_registry.json, and the
cover/results/configuration/evaluation figure set so the manuscript can be audited without reading the PDF.
The protocol emits MadlibConfig, review scenario, explicit/default path inventory, validated lexicon, digest input records, selection
invariant set, TokenPlan, enabled section set, section variables, Markdown evidence tables, claim-aligned evidence surface, registered
figure set, output/data, output/reports, and output/figures, output/manuscript, validated project output, review packet, copied
publication-review bundle, fork migration notes.
Project tests cover deterministic token choice, seed sensitivity, category-input
sensitivity, malformed config rejection, section disablement, artifact writing, and unresolved manuscript-token detection. The shared
output validator then checks rendered PDFs, Markdown, figure registry, evidence registry, and design overlays.
The copied root output is therefore a consequence of local source and config. Generated files remain disposable; the durable contract
is the ability to regenerate them from the tracked project tree and to observe the same validation gates passing.
• Config hash: cc5f4a752d0b55bc
• Generated: 2026-06-26T13:42:19Z
• Python: 3.12.13
• Platform: Darwin arm64

## Page 31

9
Limitations: Non-Claims, Misuse Modes, and Human Review
The limitations section is configured to state non-claims, identify misuse modes, preserve human review, and require domain validators
for domain claims. The central limitation is that deterministic token injection can make manuscript assembly auditable, but it cannot
make a weak claim true or a missing source appear.
The declared failure modes are Unresolved placeholder, Overclaimed generated prose, Config-source drift, Figure provenance gap,
Domain misuse, Method row drift, Field-origin opacity, Visual-method mismatch, Fork without validators, Digest invariant drift,
Claim ledger omission, Review packet incompleteness, Fork migration ambiguity. They are included in the manuscript because this
template is meant to teach the boundary, not hide it. A useful fork should extend this table when it adds domain-specific claims,
validators, or publication targets.
The author remains responsible for theory, citations, reader expectations, and domain evidence. The generator can enforce structure
and provenance; it cannot supply judgment. That division is the main safety property of the exemplar.
9.1
Failure Modes
Failure mode
Risk
Detection
Mitigation
Unresolved placeholder
A source Markdown token is
added without a generated
variable.
Project test scans
output/manuscript and
rendered web output for
uppercase placeholders.
Add variable generation, tests,
and rerun
z_generate_manuscript_variables.p
before render.
Overclaimed generated prose
Generated text implies a
standalone DOI, empirical
validation, or external release
that does not exist.
Claim ledger review,
publication metadata review,
and evidence registry
validation.
Keep publication fields blank
and use local-only claim
boundaries until evidence
exists.
Config-source drift
Documentation names a
schema feature that source
code no longer parses.
Config tests instantiate the
schema and docs point to
generated tables.
Update config loader, docs,
and tests together.
Figure provenance gap
A figure is referenced in
manuscript output without a
registry entry.
Stage 04 figure registry
validation.
Emit figure_registry.json from
src.analysis before rendering.
Domain misuse
A fork treats lexical
substitution as evidence for a
domain-specific research claim.
Failure-mode table, scope text,
and downstream domain
validators.
Add domain-specific data,
validators, and claim ledgers
before making domain claims.
Method row drift
The Methods prose describes a
protocol row or phase that no
longer exists in config.
Composition tests, summary
report review, and generated
method tables.
Update method_protocol,
pipeline_phases, composition
prose, and tests together.
Field-origin opacity
A rendered field appears
configured even though it
came from a loader default.
configured_field_inventory.json
and configured-field summary
figures.
Expose explicit/default counts
in the manuscript and review
optional defaults before
release.
Visual-method mismatch
A figure implies a method
claim that is not backed by
generated data or registry
metadata.
Figure registry validation,
nonblank figure tests, and
manual visual QA.
Generate figures only from
config, TokenPlan, and
inventory data, then rerun
validation.
Fork without validators
A downstream project changes
vocabulary or claims but keeps
only the exemplar’s generic
gates.
Authoring contract review and
claim ledger review.
Add domain validators,
domain evidence artifacts, and
claim-ledger entries before
asserting domain findings.
Digest invariant drift
A future edit lets renderer
state, file order, or ambient
prose influence token selection.
Seed-stability,
category-sensitivity, and
method-invariant tests.
Keep token selection isolated
in src/tokens.py and document
allowed digest inputs in
Methods.
Claim ledger omission
A generated claim appears in
prose or documentation
without a matching local
evidence row or non-claim
boundary.
Claim ledger review, evidence
registry validation, and
documentation review.
Add claim-ledger rows or
remove the unsupported claim
before rendering copied
outputs.

## Page 32

Failure mode
Risk
Detection
Mitigation
Review packet incompleteness
A reviewer receives PDF or
HTML without the data,
reports, figures, validation
output, or copy statistics
needed to inspect the method.
Stage 05 output statistics and
copied-output validation.
Regenerate Stages 02-05 and
include the full copied output
surface in review.
Fork migration ambiguity
A fork leaves authors unsure
which exemplar surfaces must
change for a domain-specific
manuscript.
README, STANDALONE
notes, manuscript README,
and Authoring Contract
review.
Document required config,
source, test, claim-ledger,
validator, and pipeline changes
before making domain claims.

## Page 33

10
Scope: Related Generators and Responsible Forking
The exemplar is a pipeline testbed, not a natural-language quality benchmark. It covers deterministic token selection, conditional
section bodies, section-title injection, provenance tables, and generated evidence artifacts. It does not evaluate semantic originality,
factual truth beyond the local configuration, or reader preference.
The configured scope moves are distinguish generation from truth, limit publication claims, point to local evidence, and explain respon-
sible forking. The closest related systems are not general-purpose chatbots but source-owned report generators, literate-programming
documents, static-site data templates, and reproducible manuscript pipelines. template_madlib contributes a deliberately small version
of that idea for research manuscripts.
Publication metadata remains conservative. The local CITATION.cff, .zenodo.json, and codemeta.json describe this exemplar inside
the shared template repository; they do not claim a live standalone DOI, external release, or empirical validation outside the generated
local artifacts.

## Page 34

11
Authoring Contract: Human Review and Forking Obligations
The authoring contract is configured to state human responsibilities, name fork obligations, connect review to generated evidence,
require domain validators before domain claims, and document fork migration notes.
It addresses pipeline maintainers directly
because the generator can preserve structure, provenance, and reviewability, but it cannot decide what a field should claim. Human
authors remain responsible for theory, interpretation, citations, and the choice to publish.
The declared obligations are Review generated claims, Review config diffs, Extend claim evidence, Add domain validators, Rerun the
full project path, Review method invariants, Assemble reviewer packet, Write fork migration notes. They convert responsible use into
a checklist: review generated claims in the hydrated manuscript, inspect the generated evidence tables, extend the claim ledger when
new claims appear, and add domain-specific validators before the template is used for a real empirical or theoretical manuscript.
The quality standard is claim humility. A fork that only changes words has not preserved the exemplar. A responsible fork changes
the config, adjusts source-owned composition where necessary, regenerates artifacts, and reruns the same tests and validation gates
before treating the output as reader-ready.
11.1
Authoring Obligations
Obligation
Required action
Review surface
Review generated claims
Inspect hydrated manuscript bodies
before copied outputs are treated as
reader-ready.
output/manuscript and output/web
Review config diffs
Treat lexicon, slot, title, move, and
section-switch edits as source-data
changes.
manuscript/config.yaml
Extend claim evidence
Update the claim ledger when generated
prose adds a new claim boundary.
data/claim_ledger.yaml
Add domain validators
Add tests and validation artifacts before
using the template for domain-specific
claims.
tests and output/reports
Rerun the full project path
Regenerate analysis artifacts, render
outputs, validate outputs, and copy
deliverables.
pipeline command logs
Review method invariants
Confirm changed tokens are explained
only by seed, slot, category, ordinal, or
category inventory changes.
src/tokens.py, token_inventory.json, an
d output/manuscript/02_methodology.md
Assemble reviewer packet
Provide hydrated manuscript, rendered
outputs, figures, data, reports, validation
report, and output statistics together.
output/templates/template_madlib
Write fork migration notes
Document config, source, test, validator,
and claim-ledger changes required by a
domain fork.
README.md, STANDALONE.md, and data/clai
m_ledger.yaml

## Page 35

References


---
*Extraction method: pymupdf*