agentleFS
slogsdonslogsdon/skills-workflows/skills/workflow-monthly-skill-audit/SKILL.md

workflow-monthly-skill-audit

Skill0 starsChanged 3 months agoMITView on GitHub

Run the monthly Claude Code skill hygiene pipeline — skill-audit (rank skills by usage + quality) → consolidate-memory (prune stale memories) → vault-lint (catch orphans created by deprecated workflows). Use when the 1st of the month rolls around, when the user says "monthly skill audit", "is my context layer healthy", "skill hygiene check", or after deprecating a workflow / removing a skill. Do NOT use for single-skill review (read the SKILL.md directly) or in the same week as the last audit. Output goes to ~/Downloads + Inbox/.

The file

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 2 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    After deprecating a workflow or removing skills, to catch downstream effects.

    line 13
  2. Distinctive:

    When the user explicitly asks "is my skill layer healthy".

    line 14
  3. Very specific:

    Default: run audit → memory → vault-lint in order, with checkpoints between each.

    line 51
  4. Very specific:

    If <step> is not in {audit, memory, vault-lint}, print:

    line 55
  5. Very specific:

    If a stage errors, stop and print: which step failed, the underlying skill's error, and the resume command (/workflow-monthly-skill-audit …

    line 83

≈ “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

Skill: /workflow-monthly-skill-audit

Three-stage hygiene pipeline. Mirrors the message of the LLM-context-files post: context decays without maintenance.

When to use

  • 1st of every month (cron candidate — pair with /loop or /schedule).
  • After deprecating a workflow or removing skills, to catch downstream effects.
  • When the user explicitly asks "is my skill layer healthy".

When NOT to use

  • Single-skill review → read the SKILL.md directly.
  • Inside the same week as the last audit — usage data won't have shifted.

Steps

Step nameStageInvokes
auditSkill auditskill-audit
memoryMemory consolidationanthropic-skills:consolidate-memory
vault-lintVault lintvault-lint

Stage 1: Skill audit — step audit — invokes skill-audit

Produces ~/Downloads/skill-audit-report.md.

Output: report file exists. Checkpoint: review top/bottom 5 before pruning memory.

Stage 2: Memory consolidation — step memory — invokes anthropic-skills:consolidate-memory

Reflective pass over MEMORY.md; merges duplicates, prunes stale facts, fixes the index.

Output: MEMORY.md updated; commit summary in conversation. Checkpoint: confirm memory edits look correct.

Stage 3: Vault lint — step vault-lint — invokes vault-lint

Catches orphan notes / broken links that may have been left by deprecated workflows.

Output: lint report in conversation.

Resume protocol

Default: run audit → memory → vault-lint in order, with checkpoints between each.

--resume-from <step>: skip every step before <step> and begin from <step>. The named step itself runs.

If <step> is not in {audit, memory, vault-lint}, print:

Invalid resume-from step: <step>
Valid steps: audit, memory, vault-lint

…and stop.

Required state when resuming

  • --resume-from audit — same as a fresh run.
  • --resume-from memory — consolidate-memory operates on MEMORY.md directly; it does not need the audit report. If the audit report is absent, warn the user that the typical flow uses audit findings to inform memory pruning, and ask whether to continue.
  • --resume-from vault-lint — independent of prior steps; runs against the live vault.

Examples:

/workflow-monthly-skill-audit --resume-from memory     # audit report from a prior session
/workflow-monthly-skill-audit --resume-from vault-lint # just want the orphan check today

Other flags

  • --no-pause for cron-driven runs.

Output destination

Stage 1 report goes to ~/Downloads/skill-audit-report.md and is auto-copied to vault Inbox/Skill Audit YYYY-MM-DD.md if the user opts in (asked at Stage 1 checkpoint).

Failure handling

If a stage errors, stop and print: which step failed, the underlying skill's error, and the resume command (/workflow-monthly-skill-audit --resume-from <step>).

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.