agentleFS
MicrosoftKnown publishermicrosoft/vscode-team-kit/github-inbox/skills/inbox-memory/SKILL.md

inbox-memory

Skill65 starsChanged 16 days agoMITView on GitHub

Read and write agent storage — memory and rules with environment-aware fallback

The file

We split this skill into the rules it gives an agent

11 rules. About 0 are common, 2 are distinctive, and 9 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.

  • 9

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Check for existing CLI memory: see if ~/.copilot/github-inbox-memory.md or ~/.copilot/github-inbox-rules.md exist and are non-empty (use ls …

    line 34
  2. Very specific:

    If #memory is unavailable → use file system (~/.copilot/).

    line 36
  3. Very specific:

    If #memory is available AND ~/.copilot/ has NO existing inbox files → use #memory.

    line 37
  4. Very specific:

    If #memory is available AND ~/.copilot/ has existing inbox files → ASK the user with #askQuestions which backend to use.

    line 38
  5. Very specific:

    Read ~/.copilot/github-inbox-memory.md or ~/.copilot/github-inbox-rules.md using #read (VS Code) or #view (CLI).

    line 53
  6. Very specific:

    If the file doesn't exist yet, skip the delete and just use create.

    line 62
  7. Very specific:

    Ensure ~/.copilot/ exists: mkdir -p ~/.copilot

    line 65
  8. Very specific:

    Write the file using #edit or via terminal: write full content to ~/.copilot/github-inbox-memory.md.

    line 66
  9. Distinctive:

    Always read before writing to avoid losing existing content

    line 70
  10. Distinctive:

    Keep memory under 100 lines — deduplicate, keep last 5 session logs, be concise

    line 72
  11. Very specific:

    NEVER use repo memory (/memories/repo/) — only the files listed above

    line 73

≈ “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

Inbox Memory

This skill reads and writes persistent memory and rules files for the Inbox agent. It abstracts storage so all agents use a single skill instead of calling #memory directly.

File names

  • Memory: github-inbox-memory.md — session context, patterns, preferences
  • Rules: github-inbox-rules.md — user's rules and preferences

Storage strategy

Two possible backends:

  • #memory tool (VS Code Chat) — stores under the virtual path /memories/
  • File system (~/.copilot/) — used by Copilot CLI; also usable from VS Code Chat
    • ~/.copilot/github-inbox-memory.md
    • ~/.copilot/github-inbox-rules.md

How to pick a backend (run ONCE per session, cache the choice)

  1. Detect availability of #memory: try { "command": "view", "path": "/memories/github-inbox-memory.md" }. If the tool itself errors as unavailable, #memory is NOT available.
  2. Check for existing CLI memory: see if ~/.copilot/github-inbox-memory.md or ~/.copilot/github-inbox-rules.md exist and are non-empty (use ls -la ~/.copilot/ or #view).
  3. Decide:
    • If #memory is unavailable → use file system (~/.copilot/). No prompt.
    • If #memory is available AND ~/.copilot/ has NO existing inbox files → use #memory. No prompt.
    • If #memory is available AND ~/.copilot/ has existing inbox files → ASK the user with #askQuestions which backend to use. Show them what was found (file names + sizes or a one-line preview). Options:
      • "Use ~/.copilot/ (existing CLI memory)" — recommended when CLI files exist
      • "Use VS Code #memory (start fresh / keep VS Code-only)"
      • "Use VS Code #memory and import from ~/.copilot/" — copy current CLI files into #memory once
  4. Cache the chosen backend for the rest of the session and use it for ALL subsequent reads and writes. Do NOT mix backends within a session.

Reading

Via #memory:

{ "command": "view", "path": "/memories/github-inbox-memory.md" }
{ "command": "view", "path": "/memories/github-inbox-rules.md" }

Via file system (fallback): Read ~/.copilot/github-inbox-memory.md or ~/.copilot/github-inbox-rules.md using #read (VS Code) or #view (CLI). If the file doesn't exist, treat as empty / first-time user.

Writing

Via #memory:

{ "command": "delete", "path": "/memories/github-inbox-memory.md" }
{ "command": "create", "path": "/memories/github-inbox-memory.md", "file_text": "<content>" }

If the file doesn't exist yet, skip the delete and just use create.

Via file system (fallback):

  1. Ensure ~/.copilot/ exists: mkdir -p ~/.copilot
  2. Write the file using #edit or via terminal: write full content to ~/.copilot/github-inbox-memory.md.

Rules

  • Always read before writing to avoid losing existing content
  • Memory and rules are separate files — don't mix them
  • Keep memory under 100 lines — deduplicate, keep last 5 session logs, be concise
  • NEVER use repo memory (/memories/repo/) — only the files listed above

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.