agentleFS
n-n-coden-n-code/n-n-code-skills/.agents/skills/documenter-coauthoring/SKILL.md

documenter-coauthoring

Skill4 starsChanged 11 days agoMITView on GitHub

Companion overlay to documenter for user-directed, multi-round coauthoring of specs, PRDs, proposals, ADRs, decision docs, and similar technical documents. Use when the user wants to shape or approve structure, set checkpoints, iterate section by section, or preserve decisions across rounds. Document size neither selects nor excludes it. Do not use alone or for autonomous one-pass drafting or review.

The file

We split this skill into the rules it gives an agent

9 rules. About 0 are common, 8 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈8

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Add this to documenter; it changes collaboration, not documentation rules.

    line 8
  2. Distinctive:

    read supplied files and existing docs before asking discoverable questions

    line 13
  3. Distinctive:

    ask one batched set of questions only for unknowns that materially affect the next checkpoint

    line 16
  4. Distinctive:

    treat delegation as authority to skip covered pauses, not as authority to hide ripple effects, weaken evidence or safety checks, or cause …

    line 32
  5. Distinctive:

    write summaries after the decision and evidence are stable, then place them where readers need them

    line 41
  6. Distinctive:

    treat section-specific feedback as local.

    line 44
  7. Distinctive:

    keep one canonical artifact rather than parallel conflicting drafts

    line 51
  8. Distinctive:

    before declaring completion, perform a cold-read pass against the named audience and task without relying on conversational context.

    line 58
  9. Distinctive:

    when the user pauses or requests a handoff, return the working state needed to resume and do not claim the document is complete

    line 64

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Documenter Coauthoring

Add this to documenter; it changes collaboration, not documentation rules.

Workflow

  1. Establish the working state:
    • read supplied files and existing docs before asking discoverable questions
    • infer the audience, outcome, template, constraints, and output target
    • invite an unstructured context dump when useful
    • ask one batched set of questions only for unknowns that materially affect the next checkpoint
    • track the accepted outline and sections, decisions, general preferences, assumptions, open questions, delegated authority and its limits, checkpoint cadence, and next action
    • on resume, or before switching from chat-only draft to applied edits, refresh the canonical artifact, relevant evidence, repository state, and initial diff; mark affected accepted sections stale when material evidence or decisions changed
  2. Agree on structure and user control:
    • confirm and reuse an established structure; when structure is unresolved, propose an outline with material alternatives, assumptions, and a recommendation
    • derive checkpoint cadence from the request; if it is unspecified, propose the smallest useful next checkpoint
    • let the user approve, revise, waive, or delegate checkpoints and decisions
    • treat delegation as authority to skip covered pauses, not as authority to hide ripple effects, weaken evidence or safety checks, or cause external side effects
    • pause only at an agreed checkpoint or for a material choice outside delegated authority; do not add gates because the document is large or uncertain
  3. Draft the agreed section or logical tranche:
    • work in dependency order; prototype a high-uncertainty section first only when it could change the outline
    • write summaries after the decision and evidence are stable, then place them where readers need them
    • preserve accepted material and decisions
    • treat section-specific feedback as local; carry clearly general preferences such as tone, terminology, or audience forward, and ask only when their scope is materially ambiguous
    • surface ripple effects before changing an accepted decision; when delegated authority covers the change, disclose them without adding a checkpoint and mark affected accepted sections stale until reconciled
  4. Reconcile at each checkpoint:
    • keep one canonical artifact rather than parallel conflicting drafts
    • show the current tranche or changed blocks during iteration while keeping the complete artifact coherent at milestones and completion
    • report what changed, current decisions and assumptions, unresolved material questions, validation state, and the next proposed action or gate
  5. Finish or hand off:
    • apply the baseline validation rules
    • before declaring completion, perform a cold-read pass against the named audience and task without relying on conversational context; report it as self-review, not an independent fresh-reader test
    • declare completion only when the baseline completion criteria and cold-read pass are satisfied; seek approval at an agreed completion checkpoint, or complete autonomously within delegated authority
    • when the user pauses or requests a handoff, return the working state needed to resume and do not claim the document is complete

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.