agentleFS
Sign inStart free
arbelharbelh/perdure/skills/raw/SKILL.md

raw

Skill0 starsChanged 28 days agoMITView on GitHub

Sweep or verify the session index at perdure/raw/INDEX.jsonl

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 2 are distinctive, and 0 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.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Run the sweep (safe, idempotent, additive):

    line 18
  2. Distinctive:

    No transcript text, no file paths, no prose — only ids, counts, hashes, timestamps and low-cardinality labels (branch, tool version, …

    line 47

≈ “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

The session index

Reads Claude Code's session store under ~/.claude/projects/.

perdure/raw/INDEX.jsonl is the population: one line per session that ran in this project. It is the only artifact here derived from what actually happened rather than from what someone chose to write down.

Run the sweep (safe, idempotent, additive):

python3 "${CLAUDE_PLUGIN_ROOT}/scripts/raw-index.py" --root .

Without ${CLAUDE_PLUGIN_ROOT}, the script sits in the plugin checkout's root scripts/ directory (Codex keeps the whole checkout), or in scripts/ beside this SKILL.md (skills CLI).

Other modes:

python3 "${CLAUDE_PLUGIN_ROOT}/scripts/raw-index.py" --root . --summary   # print the population, write nothing
python3 "${CLAUDE_PLUGIN_ROOT}/scripts/raw-index.py" --root . --verify    # re-derive and compare; exit 1 on mismatch

Then:

  1. Report the population, not just the file count — sessions present, sessions pruned, prompts, subagent transcripts.
  2. If any session is marked pruned, say so plainly. Its transcript is gone from the local store; the entry and its digest remain, so a record can still cite it, but the payload cannot be re-read. An honest gap is a record.
  3. Never edit this file by hand. It is swept, and --verify will fail on a hand edit.

What it deliberately does not contain

No transcript text, no file paths, no prose — only ids, counts, hashes, timestamps and low-cardinality labels (branch, tool version, entrypoint, model). That is why it is committed while the transcripts it points at are not: the safety is a property of the schema, not of a scan that might miss something. The sweep refuses to write an entry that would violate it.

What it is for

The 2026-08-31 lifecycle dry run produced four workflow records for five units of work, and nothing in the system could tell — perdure knew only about records someone chose to write. This index is what makes that kind of gap countable: it gives the denominator. It does not decide whether a session should have produced a record; that judgement belongs to whatever reads this file.

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.

Check your setup, free