agentleFS
epechuzalepechuzal/claude-skill-recall/session-recall/SKILL.md

session-recall

Skill1 starsChanged 7 months agoMITView on GitHub

Use when the user references a past conversation ("we talked about this", "remember when", "what did you recommend", "didn't we discuss"), or when a topic surfaces that may have a saved session note. Auto-load the session index and surface relevant notes naturally.

5 other projects use a version of this.

Adopted by

Most recently taken up by

  • Other versions5 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 5

    1. iouovonoi/session-recall★ 4same name
    2. kratocz/skills★ 1same name
    3. Yaagik/session-recall★ 1same name
    4. CavsSatyamKhatri/claude-session-recall★ 0same name
    5. SoyErnoModo/session-recall★ 0same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repositoryOne other file it gives its agents.

The file

We split this skill into the rules it gives an agent

5 rules. 0 are common, 4 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 4 edited versions we read keep these.

  • 4

    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:

    Before giving advice that might contradict a past session's ## Key Decisions, check the index.

    line 39
  2. Distinctive:

    Read the index file: when a relevant trigger fires — don't load it preemptively, wait for the trigger.

    line 45
  3. Distinctive:

    Read the matched session file: Don't summarize from memory.

    line 47
  4. Distinctive:

    Don't be aggressive. Surface only when genuinely relevant.

    line 54
  5. Distinctive:

    Always log — even when the scan finds nothing.

    line 76

“Common” is measured: kept by at least 6 in 10 of the 4 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

Session Recall

Surface relevant past conversation notes when topics from prior sessions resurface.

This is an ambient skill — no slash command. Fires when the conversation context matches a past session.


Setup

Two paths to configure (same as the tldr skill — share them):

  • Index file — one row per saved session: date, title, tags (default: sessions/INDEX.md)
  • Sessions directory — where session notes live (default: sessions/)

Companion: the tldr skill writes the notes that this skill reads.


When to Surface

Explicit recall

User says: "we talked about this", "remember when", "what did you recommend", "didn't we discuss X".

→ Search the index for matching tags or title keywords. Read the session file. Reference it naturally.

Implicit recall

User mentions a topic that has a saved session.

→ Check the index. If the past session is genuinely relevant, surface it: "We covered this on [date] — [link]. Want me to pull up the details?"

Consistency check

Before giving advice that might contradict a past session's ## Key Decisions, check the index. If a relevant session exists, flag it: "Note: in our [date] session, we decided X. Want to revisit that?"


How to Use

  1. Read the index file when a relevant trigger fires — don't load it preemptively, wait for the trigger.
  2. Search by tag or title keyword. Index is reverse chronological; scan top entries first.
  3. Read the matched session file. Don't summarize from memory.
  4. Reference it naturally in the response, with a markdown link to the session file.

Restraint

Don't be aggressive. Surface only when genuinely relevant. Three failure modes to avoid:

  • Forced recall — dragging in a vaguely-related session because it shares a tag. If the past session won't change the current answer, don't mention it.
  • Repetition — if the user already saw the recall in this conversation, don't keep re-surfacing it.
  • False memory — never claim "we discussed this" without reading the actual session file. Recall the contents, not your impression of them.

But never let the user ask "didn't we talk about this?" when the answer is right there in the index. That's the failure to avoid on the other side.


Optional: Activation Logging

For tuning false-positive vs. true-positive ratios, log every fire (including no-match):

mkdir -p .claude-recall-log && printf '%s\n' "$(jq -nc \
  --arg ts "$(date -Iseconds)" \
  --arg trigger "<short paraphrase of what user said>" \
  --arg matched "<session slug or 'no-match'>" \
  '{ts: $ts, trigger: $trigger, matched: $matched}')" >> .claude-recall-log/log.jsonl

Always log — even when the scan finds nothing. False positives (skill firing when it shouldn't) only show up in the no-match rows. Most users won't need this; it's for tuning the skill against a real corpus over time.


Companion Command

The tldr skill writes new session notes. This one reads them. Both touch the same index.

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.