agentleFS
UnfoldCRO-OrganizationUnfoldCRO-Organization/cro-agent-skills/skills/update-md/SKILL.md

update-md

Skill0 starsChanged 22 days agoMITView on GitHub

Update project Markdown documentation from actual progress when the user says update md, update docs, save progress, update tasks, update context, update project md, record today's work or write progress.

The file

Show its rules and text

We split this skill into the rules it gives an agent

20 rules. About 0 are common, 16 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈16

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Update only relevant project Markdown documentation using actual completed work, actual changed files and actual verification results.

    line 13
  2. Distinctive:

    Use this workflow when the user says:

    line 17
  3. Distinctive:

    update md / update docs / save progress / update tasks

    line 18
  4. Distinctive:

    update context / update project md / record today's work

    line 19
  5. Distinctive:

    write progress / update markdown / save current status

    line 20
  6. Very specific:

    Do not use task-dispatch.

    line 26
  7. Distinctive:

    Do not invoke agents, MCP servers, plugins, or unrelated skills.

    line 27
  8. Distinctive:

    Do not perform new application work unless separately requested.

    line 28
  9. Very specific:

    Read the project index (CLAUDE.md, or .claude/CLAUDE.md in a repo with a remote) only if not already read in current context.

    line 36
  10. Distinctive:

    Read only the documentation files needed for the update.

    line 44
  11. Very specific:

    Never recreate legacy files (tasks.md, context.md, memory.md, rule.md, docs/deploy.md…) in a v2 project.

    line 73
  12. Distinctive:

    Do not write assumed completion.

    line 96
  13. Distinctive:

    Do not mark tasks complete without implementation evidence.

    line 97
  14. Distinctive:

    Do not claim tests passed unless performed.

    line 98
  15. Distinctive:

    Do not write planned testing as completed testing.

    line 99
  16. Distinctive:

    Do not write long narrative progress summaries.

    line 100
  17. Very specific:

    Do not place ordinary progress inside CLAUDE.md, rule.md or docs/06-AGENTS.md.

    line 101
  18. Distinctive:

    Do not update unrelated project documentation.

    line 102
  19. Distinctive:

    Do not copy private cross-project data into Markdown.

    line 103
  20. Distinctive:

    Do not record push, tag, release or deployment unless it actually happened.

    line 104

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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

Update Project Markdown from Actual Progress

Request details: $ARGUMENTS

Purpose

Update only relevant project Markdown documentation using actual completed work, actual changed files and actual verification results.

Activation Phrases

Use this workflow when the user says:

  • update md / update docs / save progress / update tasks
  • update context / update project md / record today's work
  • write progress / update markdown / save current status

Mandatory Restriction

For this request:

  • Use update-md only.
  • Do not use task-dispatch.
  • Do not invoke agents, MCP servers, plugins, or unrelated skills.
  • Do not perform new application work unless separately requested.
  • Project docs stay on this machine: never stage, commit or push CLAUDE.md, .claude/CLAUDE.md, AGENTS.md, AGENTS.local.md, docs/NN-*.md, docs/*.local.md, docs/ref/ or .claude/skills/ in a repo with a remote (there the index is .claude/CLAUDE.md). If git status lists one, add the exclude block (~/.claude/references/project-docs-standard.md §7) to .git/info/exclude instead.

Mandatory Inspection Before Writing

Before updating Markdown:

  1. Identify the active project root.
  2. Read the project index (CLAUDE.md, or .claude/CLAUDE.md in a repo with a remote) only if not already read in current context.
  3. Inspect actual current progress using:
    • Work completed in the current session.
    • Files changed in the current session.
    • git status --short
    • git diff --stat
    • Relevant git diff output only where needed.
    • Build, lint, test, browser or manual verification actually performed.
  4. Read only the documentation files needed for the update.

File Selection Rules

v2 projects (the index has ## Docs map)

Actual SituationFile to Update
Task added, started, blocked, waiting or phase changeddocs/08-IMPLEMENTATION.md
Session checkpoint (done · next · open decisions)docs/08-IMPLEMENTATION.local.md §Checkpoint (overwrite, ≤10 lines)
Merged, shipped or releaseddocs/17-CHANGELOG.md (newest first) — and remove the item from 08
Approved scope changed or deliverable status changeddocs/01-PRD.md §Scope / §Scope status
UI or design decisiondocs/02-DESIGN.md
Structure or technical decisiondocs/03-ARCHITECTURE.md (§Decisions: - YYYY-MM-DD — decision — why)
Schema or migration processdocs/04-DATABASE.md
Endpoint, server action or webhookdocs/05-API.md
How agents should work on this project, project skill addeddocs/06-AGENTS.md
Client, URL, environment or access pointerdocs/07-README.md (sensitive → .local.md)
Setup steps or env var namesdocs/09-SETUP.md
Feature specced, built or retireddocs/10-FEATURES.md
Auth, roles or security ruledocs/11-AUTH-SECURITY.md
MCP / AI tooling for this projectdocs/12-MCP.md
Third-party service or tracking IDdocs/13-INTEGRATIONS.md
Test setup or QA checklistdocs/14-TESTING-QA.md
Reusable debugging lessondocs/15-DEBUGGING.md (### symptom · Cause · Fix)
Deployment process or release policydocs/16-DEPLOYMENT.md
Hard rule, stack, commands, or a slot createdthe index — CLAUDE.md / .claude/CLAUDE.md (keep ≤70 lines; update the Docs map)
  • Slot marked unwritten and the work touched that area: create it now — facts from the code only, each section ending Source: `path` (standard: ~/.claude/references/project-docs-standard.md). Remove the unwritten marker.
  • Never recreate legacy files (tasks.md, context.md, memory.md, rule.md, docs/deploy.md…) in a v2 project.

Legacy projects (no Docs map)

Actual SituationFile to Update
Active task changed, completed or replaceddocs/task.md
New pending item, blocker or approval waitdocs/pending.md
Approved scope changeddocs/scope.md
Proposal or commercial commitment formally changeddocs/proposal.md
Approved deliverable completion status changeddocs/completed-vs-scope.md
Meaningful completed implementation changedocs/changelog.md
Deployment process or release policy changeddocs/deploy.md
Confirmed release, push, tag or deployment completeddocs/changelog.md
Meaningful technical implementation state changedcontext.md
Reusable UI or design decision changedDESIGN.md
Reusable debugging lesson foundmemory.md
Architecture or specification-level rule changedspec.md
Project operating instructions changedthe index (CLAUDE.md / .claude/CLAUDE.md)
Permanent unique project hard rule changedrule.md

Never Do This

  • Do not write assumed completion.
  • Do not mark tasks complete without implementation evidence.
  • Do not claim tests passed unless performed.
  • Do not write planned testing as completed testing.
  • Do not write long narrative progress summaries.
  • Do not place ordinary progress inside CLAUDE.md, rule.md or docs/06-AGENTS.md.
  • Do not update unrelated project documentation.
  • Do not copy private cross-project data into Markdown.
  • Do not record push, tag, release or deployment unless it actually happened.

Required Progress Format

## YYYY-MM-DD . <Short Task Name>

- Done: <Actual completed work. Maximum 2 lines.>
- Test: <Actual verification completed, or `Not tested yet`. Maximum 2 lines.>
- Next: <Pending step only when required. Maximum 2 lines.>

For several related completed items:

## YYYY-MM-DD . Progress Update

- Done: <First actual completed item. Maximum 2 lines.>
- Done: <Second actual completed item. Maximum 2 lines.>
- Test: <Actual verification completed, or `Not tested yet`. Maximum 2 lines.>
- Next: <Pending step only when required. Maximum 2 lines.>

Changelog Format

## YYYY-MM-DD . <Short Change Title>

- Changed: <Confirmed completed change. Maximum 2 lines.>
- Test: <Actual verification performed, or `Not tested yet`. Maximum 2 lines.>
- Reference: <Scope item or affected feature only when useful. Maximum 1 line.>

Release Changelog Format

## YYYY-MM-DD . Release <version-or-short-title>

- Changed: <Confirmed released change. Maximum 2 lines.>
- Test: <Actual deployment or post-release verification. Maximum 2 lines.>
- Release: <Actual branch and tag only when used. Maximum 1 line.>

Completion Response

Respond only with:

Updated:
- <file path>: <short description>

Recorded:
- Done: <one line>
- Test: <one line>
- Next: <one line only when applicable>

Git Summary:
- Modified documentation files: <file paths>
- Application source files modified: <None, unless separate user-requested implementation>
- Uncommitted changes detected: <yes/no based on actual git status>

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.