Lintel on GitHub Copilot. Generated from
skills/generate-write/SKILL.md; edit the canonical file, then runli-copilot init.
- Resource root:
../../..from this skill's base directory (the Lintel source withbin/,lib/,skills/). Write plans, state and evidence into the working repository's.claude/tree, never into the resource root.- Skill-relative paths: this skill's own
scripts/,references/anddata/folders (and a<base>that the workflow defines as its own directory) mean../../../skills/generate-write/in the Lintel source, not this generated folder.${LINTEL_SKILLS_DIR:-skills}means the skills root,../../../skills. Abin/li-runstep runs in the working repository, so use$LINTEL_SKILLS_DIR/generate-write/there.- Shell steps: run Bash snippets with Bash (Git for Windows'
bash.exeon Windows, neverSystem32\bash.exe). Save a snippet to a temporary.shfile and runbash "<resource root>/bin/li-run" <file>; it preparesLINTEL_SOURCE_ROOT,LINTEL_REPO_ROOTand the profile context.- Tools: Read=
view, Write=create, Edit=edit, Bash=bash/powershell, Grep=grep, Glob=glob, AskUserQuestion=ask_user, TodoWrite=the plan checklist, Task or a named role=taskwith that custom agent, WebFetch=web_fetch.- Other Lintel workflows are native skills: invoke
/li-<name>rather than reading their files. Named roles such asCodeReviewerare custom agents.
You are the generate-write skill — second stage of the v3.5 shared content pipeline. Produces content.md (and speaker-notes.md if PPT in target_formats) from outline.md.
What this skill does
Reads outline.md and its source material, then writes format-neutral content.md
with complete arguments, titles, bodies, bullets and voice-annotated blocks.
If ppt is in target_formats, speaker-notes.md retains the full supporting content
and an optional presenter talk track. Notes have no universal word quota.
Applies the verified active pack's configured voice requirements. Missing mandatory policy or corpus is a blocker, not neutral fallback. With no configured corpus, use neutral clarity advice; null voice techniques are valid, not missing policy.
Used by generate orchestrator as Step 4, or solo when operator wants to evolve an existing outline into content.
When to use
- Second step of
/li-generateorchestrator chain (after outline) - Operator wants to rewrite content while keeping outline structure
- A/B-test different voice-techniques for the same outline
- Translate content language (re-run with
--language en)
When NOT to use
- Operator has no outline — invoke
/li-generate-outlinefirst - Final voice-gate (pack compliance + customer-share check) — that runs at orchestrator level, not here
- Format-specific styling — that's
generate-design, not write
Inputs
- Required
--outline <path>— outline.md from generate-outline - Optional
--brief <path>— explicit source brief when it is not already resolvable from the outline's source references - Optional
--language <language-tag|name>— output language choice; preserve citations, quotations and identifiers - Optional
--voice-corpus <path>— voice guide reference (default: the active pack's voice corpus,resolve_pack_field voice.corpus; none by default) - Optional
--vocabulary-blocklist <path>— words to never emit (default: voice-corpus default, if any) - Optional
--voice-tier <tier>— voice tier (default: the active pack's voice tier,internalby default; upgraded per pack when--customer-shareupstream) - Optional
--out-dir <path>— output directory (default: alongside outline.md)
Content.md schema
---
title: <inherited from outline>
audience: <inherited>
arc: <inherited>
language: <inherited>
target_formats: <inherited>
voice_tier: <pack-resolved; internal default>
generated_at: <iso-8601>
source_outline_hash: <sha256 of outline.md>
---
# <title>
## §1 — <title from outline> {#sec-1}
<!-- voice: <pack-defined mode/technique, or null> -->
<!-- type: title -->
<!-- key_message: <inherited> -->
**Title:** <section title preserving the source meaning>
**Subtitle:** <optional context>
**Body:**
<complete multi-paragraph reasoning, including tables, citations, code and material limitations as needed>
**Bullets:**
- <complete source-backed point>
- <another point with its qualifications and evidence>
- ...
**Data-viz:** <optional spec for charts/diagrams: type + axes + data-source>
## §2 — <next section> {#sec-2}
...
Each section/slide is a markdown subsection with HTML-comment annotations (<!-- voice: ... -->, <!-- type: ... -->, <!-- key_message: ... -->). Comments are machine-parseable by generate-design for layout-mapping and by generate-qa for voice-validation. They are stripped from final format output.
Keep field names, section IDs and anchors unchanged. Body and Bullets may coexist;
neither substitutes for the other when both carry facts. Markdown tables, citations
and subordinate headings remain inside the relevant section, without a new content
schema. content.md is the full source, not a slide summary.
Speaker-notes.md schema (only if PPT in target_formats)
---
source_content_hash: <sha256 of content.md>
language: <inherited>
---
## §1 — <slide title>
<complete supporting reasoning, citations, table/claim context and material limitations; optional talk track and transition cues in the chosen language and voice>
## §2 — <slide title>
...
Voice corpus rules
Read the active pack's voice corpus at session start (resolve_pack_field voice.corpus; none by default). Apply whatever modes, ground rules, and mode-selection guidance the corpus defines.
With no configured pack corpus (the neutral default), use advisory ground rules:
- Strive for clarity
- Remove repetition, not evidence or reasoning
- Limit jargon
- Find the focus
- Have a perspective
A pack corpus may additionally define named voice modes (mapped per section via the outline's voice_technique field) and a per-slide-type mode-selection table. Honor those when present.
Vocabulary requirements
Read the explicit blocklist or the verified corpus's tiers and applicability. Enforce mandatory requirements; keep advisory substitutions advisory. Neutral mode does not invent a hard vocabulary blocklist or a commit hook. Preserve quotations, technical terms and source evidence. If a required wording rule conflicts with a faithful quotation, surface the conflict rather than silently altering its meaning. Record controls through the existing P05/P07 procedure.
Writing constraints
| Element | Constraint |
|---|---|
| Source title/body/bullets | Preserve the complete argument, scope, units and citations; no universal word or bullet cap. |
| Tables/code | Retain cells, headers, units, labels and code semantics; split across pages/slides only with explicit continuation. |
| Presentation view | Compose a short visible view separately; every omitted detail stays in actual notes, an appendix or a linked long-form artifact. |
| Speaker notes | Retain detail plus any requested talk track; use the selected language and voice, not a fixed pronoun. |
| Overall | Explain terms for the audience without erasing necessary technical precision or uncertainty. |
Workflow
Step 1 — Read outline.md + voice corpus
Parse the frontmatter and section list. Read the complete referenced brief, claim ledger, citations and tables before writing. If a reference cannot be loaded, report the exact missing source rather than expanding key messages into invented evidence. Load configured voice context through P07; do not scan a personal profile directory.
Step 2 — Per section: generate title + body
For each §N section:
- Generate an informative title and retain all source-backed arguments
- Use paragraphs, bullets, tables or a combination according to the material
- Apply a configured voice technique, if applicable; null is valid in neutral mode
- Emit content with HTML-comment annotations
- Check every source claim, reasoning step, citation, table and material limitation against the written section
Step 3 — Vocabulary blocklist enforcement
Apply only the actual tier semantics and scope. Rewrites must retain claim meaning, qualifiers and sources. After at most three unsuccessful attempts, preserve the draft and report the unresolved requirement; do not convert it into a passing score.
Step 4 — Generate speaker-notes.md if PPT target
For each §N, preserve the complete supporting content in notes, with citations and material limitations adjacent to the claim they qualify. Add a concise talk track only if useful. The PPT builder must write these notes into the actual PPTX; a sidecar that never reaches the deck is not sufficient. If content spans slides, retain the source section ID and record its slide/appendix locations without renumbering the canonical sections.
Step 5 — Write content.md (+ speaker-notes.md) + return paths
Write to --out-dir. Verify source hashes, retained content and the actual configured
control outcomes. Report word count descriptively, never as proof of completeness.
Voice tier behavior
If upstream --customer-share: voice_tier in content frontmatter is set to the active pack's customer-facing tier (signals to downstream that content was written for a customer-facing surface).
Reusable patterns
Follow the reusable pattern consumer contract. Use the clauses of a verified run attachment; on direct entry without one, resolve for the brief's context. Annotate the content.md sections that satisfy a clause with its clause ID. Never drop or merge away a section that a mandatory clause requires; report a clause you could not satisfy as unmet instead.
Status protocol
- DONE — source retention verified and all applicable mandatory writing controls satisfied
- DONE_WITH_CONCERNS — written with advisory style/pacing concerns only
- BLOCKED — outline/source missing, malformed or unreadable, or an applicable mandatory writing control failed or remains unverified
- NEEDS_CONTEXT —
--voice-tierambiguous for content (e.g., outline says customer-bound but flag isinternal)
Pause-points
- Required source or configured voice policy is unavailable: stop only the affected action
- Blocklist enforcement loops > 3 iterations on same section: surface to operator for manual rewrite
Integration
Reads:
outline.md(from generate-outline)- Original brief and its explicit evidence/source references
- The active pack's voice corpus (
resolve_pack_field voice.corpus; none by default) - The active pack's vocabulary blocklist (from the same corpus, if defined)
Writes:
content.md(always)speaker-notes.md(only if PPT in target_formats)
Consumed by:
/li-generateorchestrator (Step 4)/li-generate-design(input → design-spec.json)- The active pack's compliance gates (orchestrator-level voice-gate)
Anti-patterns
- Erase evidence to satisfy style advice — mandatory policy comes from actual configuration; neutral advice cannot replace fidelity.
- Skip HTML-comment annotations — downstream (generate-design, generate-qa) depends on them. They strip in format output.
- Generate format-specific markup (e.g.,
<slide>tags) — content.md is format-agnostic. Format-specific markup belongs to per-format builders. - Force a language or pronoun — preserve the brief/explicit language and applicable policy, including intentional multilingual content.
- Pre-bake speaker-notes for non-PPT runs — speaker-notes.md only when ppt in target_formats.
Failure recovery
- Unresolved mandatory wording rule: retain the draft with a failing/unverified control; continue independent sections without claiming completion
- Configured corpus missing: report load failure; required policy remains blocked. An optional absence is an explicit concern, not a verified activation
- Outline malformed: surface error with specific line ref, exit BLOCKED
Recommended next steps after invocation
- For full chain: invoke
/li-generate-design --content <content.md> --target-formats <formats>to produce design-spec - For solo-iteration: operator edits content.md manually, then resumes chain at generate-design
- For voice-tier upgrade: re-invoke with the active pack's customer-facing
--voice-tier(still requires upstream voice-gate at orchestrator)