agentleFS
dtiger1889-opsdtiger1889-ops/claude-harness-toolbox/skills/log/SKILL.md

log

Skill31 starsChanged 31 days agoMITView on GitHub

Records reusable facts, corrections, decisions, and rule violations in the document or ledger that owns them. Use for /log, "log this", "capture this", "write that down", "don't lose this", or "log that lapse". Complements checkpoint, which owns project resume state; does not run on every answer.

18 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

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

  • 0

    Common

    Most of the 16 edited versions we read keep these.

  • 4

    Distinctive

    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:

    Read the target's structure.

    line 15
  2. Distinctive:

    Use the project's designated durable store.

    line 19
  3. Distinctive:

    Use JSON Lines records with a Markdown twin for scope and Fix candidates.

    line 23
  4. Distinctive:

    If PowerShell or the ledger is unavailable, report the blocker and preserve a proposed record in an accessible owning document where …

    line 41

“Common” is measured: kept by at least 6 in 10 of the 16 edited versions of this file we could read. “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

Log reusable findings

  1. Extract the finding from the named passage or previous answer. Capture established facts and sources; do not re-research. If nothing is durable, report "nothing to log".
  2. Classify and route:
    • Reference fact or mechanism: the owning living document.
    • Decision: its decision record or existing decisions convention.
    • Event or resume state: checkpoint or the project's changelog procedure.
    • Skipped required skill step: a skill-lapse ledger.
    • Broken general instruction: a rule-lapse ledger.
  3. Read the target's structure. Correct erroneous text in place or add a dated entry. Name the evidence and author: user decision only for an explicit user choice, otherwise model finding or recommendation.
  4. For a cross-project finding, record it at its owner and leave a concise pointer where needed. If no owner document exists, propose an appropriate home and obtain any permission required by the project's conventions.
  5. Report what was written and its clickable path briefly.

Use the project's designated durable store. Do not create a parallel memory system or convert one-off feedback into a standing rule.

Lapse ledgers

Use JSON Lines records with a Markdown twin for scope and Fix candidates. For a new ledger, copy ledger-template.md and create an empty JSONL beside it.

Resolve <skill-dir> to the directory containing this loaded SKILL.md, including plugin installations. The bundled log_lapse.ps1 requires PowerShell. Append through it only; never edit or rebuild existing JSONL rows by hand.

General rule invocation, as one PowerShell command:

& "<skill-dir>/log_lapse.ps1" -LogFile "<ledger>.jsonl" -Rule "<broken rule>" -Session "<context>" -Consequence "<cost>" -HowCaught "<how found>" -FixStatus "<correction; author and date>" -NotRepeat

Replace -Rule with -Skill "<skill>" -Step "<skipped step>" for a skill lapse.

Before appending, search earlier rows for the same failure family. Use -RepeatOf <n> instead of -NotRepeat for a confirmed repeat. Use -NotRepeat only after checking. With neither flag, the script may detect a possible repeat and exit 3 without appending; decide whether it is a repeat, then retry once with the appropriate flag. The fuzzy hint is a fallback, not a semantic duplicate check.

The script obtains model/effort from the current Claude Code transcript when available. Other runtimes pass -Model and -Effort as caller-reported values. For a lapse in another session, use -ForeignSession -Model "<actual model>" and available effort evidence so this session is not misattributed.

Record fixed lapses too. For a repeat, update the Markdown Fix candidates entry and one owning-project open item with a suggested family-level prevention change. Do not implement a new gate merely because the script prints PROMOTE.

If PowerShell or the ledger is unavailable, report the blocker and preserve a proposed record in an accessible owning document where authorized. Do not claim a ledger append succeeded.

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.