agentleFS
lorena-bordonaba-paulorena-bordonaba-pau/youtube-agent/.claude/skills/memory-review/SKILL.md

memory-review

Skill2 starsChanged 49 days agoMITView on GitHub

Review the session for stable learnings about the channel, the voice, the formats or the way of working, and propose memory updates for the creator to approve. Run by /remember, and offered before a long session is compacted. Triggered by "remember this", "save what we learned", "update the memory", "what did we learn today". In Spanish, "acuérdate de esto", "guarda lo que hemos aprendido", "actualiza la memoria", "qué hemos aprendido hoy".

6 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

1 rule. 0 are common, 0 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 5 edited versions we read keep these.

  • 0

    Distinctive

    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:

    Find the home of each candidate: Update the file that already covers it.

    line 33

“Common” is measured: kept by at least 6 in 10 of the 5 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

Memory review

WHEN

At the end of a working session, when the creator asks to remember something, or when the session-start or pre-compaction notice suggests it. Always with the creator's approval before anything is written.

Memory is what turns a consultant into a coach, but only if what goes in is stable and true. A memory full of one-off remarks is noise the agent then obeys.

EXECUTION ORDER

  1. Read memory/MEMORY.md and the files it indexes that the session touched. Nothing is proposed without knowing what is already there.
  2. List the candidates. Go through the session and keep only what will still be true next month:

    KeepLeave out
    A preference the creator stated or corrected ("tables, not prose")What only matters to today's task
    A prediction made and, later, its outcomeFigures that a tool can measure again (store the date and the tool, not the number, unless it is an outcome)
    A format, pillar or positioning decisionGuesses about why a metric moved: correlation is not a measurement
    A voice trait backed by a verbatim quoteAnything the repo already records (tools, skills, config)
  3. Find the home of each candidate. Update the file that already covers it; create a new one only when none does, with the same front matter as the rest (name, description, metadata.type) and a line in MEMORY.md.
  4. memory/rules.md is special. A candidate goes there only if the creator explicitly imposed it in this session. Quote their words and the date. A pattern you noticed is not a rule: propose it for coach_preferences.md or the journal instead, or leave it out.
  5. Propose, do not write. Show each change as: file, the exact lines to add or replace, and why it is stable. Ask once for approval of the whole batch; the creator can drop items.
  6. Write only what was approved, then update MEMORY.md if a file was added or its status changed from not populated.

SOURCES

  • The session itself and the tool outputs in it: every figure carried into memory keeps the source label and the date the tool returned it.
  • memory/*.md: config, what was already stored.
  • Nothing in this skill calls the YouTube API or spends quota.

OUTPUT

A short list of proposed changes, grouped by file, each with its reason; then, after approval, one line per file saying what was written. If nothing stable came out of the session, say so in one line: an empty review is a valid result.

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.