agentleFS
ImYourBoyRoyImYourBoyRoy/agent-continuity-stack/skills/maintain-project-memory/SKILL.md

maintain-project-memory

Skill1 starsChanged 4 months agoMITView on GitHub

Create, audit, compact, or synchronize the required MEMORY.md and related continuity artifacts. Use after meaningful project work, before a resumable handoff, or when project memory is missing, stale, unsafe, or duplicative.

2 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

8 rules. About 0 are common, 7 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈7

    Distinctive · estimate

    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. Distinctive:

    Keep it Git-ignored and out of commits, releases, and other uploads unless the user explicitly authorizes sharing it.

    line 9
  2. Very specific:

    Keep MEMORY.md as the continuity index.

    line 22
  3. Distinctive:

    After meaningful work, synchronize completed and remaining work, important path changes, discoveries, decisions, risks, validation, and the …

    line 50
  4. Distinctive:

    Never record credentials, tokens, private keys, secret values, personal data, or sensitive production content.

    line 56
  5. Distinctive:

    Avoid absolute personal paths, private hosts, device IDs, and environment dumps.

    line 58
  6. Distinctive:

    Do not claim a model's private memory or chat history is an authoritative project source.

    line 60
  7. Distinctive:

    Keep shared durable guidance in version-controlled, human-reviewable artifacts.

    line 62
  8. Distinctive:

    Before finishing substantial work, make sure another contributor can identify:

    line 69

≈ “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

Maintain Project Memory

Maintain a concise, current local MEMORY.md for every maintained project. Keep it Git-ignored and out of commits, releases, and other uploads unless the user explicitly authorizes sharing it. Use it to route future work without treating it as stronger evidence than source, Git, manifests, tests, builds, or verified runtime behavior.

Establish the foundation

Resolve the repository root, read project instructions and relevant README sections, inspect Git status, and locate MEMORY.md. If it is absent, create it unless the current request is strictly read-only or the repository forbids the change; in those cases, report the missing required foundation. Ensure MEMORY.md is ignored by Git before writing private continuity state.

Keep MEMORY.md as the continuity index. Delegate deep or specialized content:

  • issues or TASKS.md own large backlogs and acceptance criteria
  • ADRs or ARCHITECTURE.md own deep durable design decisions
  • runbooks own repeatable operations
  • README.md owns stable usage and onboarding facts
  • CHANGELOG.md owns release history
  • temporary handoffs own short-lived execution state

Link or summarize those authorities without copying them. Read references/artifact-selection.md when ownership is unclear.

Synchronize compactly

Include only applicable:

  • project snapshot and useful working-directory map
  • current goals, active work, and exact next action
  • architecture notes, decisions, and conventions
  • verified resources and dated software-version findings
  • important commands and workflows
  • known issues, risks, and blockers
  • recent material changes
  • validation performed and its environment

Label observations, inferences, and proposals. Timestamp volatile external facts. Replace stale status instead of appending an endless transcript.

After meaningful work, synchronize completed and remaining work, important path changes, discoveries, decisions, risks, validation, and the next recommended action. Do not update memory merely to restate a trivial conversation.

Safety and privacy

  • Never record credentials, tokens, private keys, secret values, personal data, or sensitive production content.
  • Avoid absolute personal paths, private hosts, device IDs, and environment dumps.
  • Do not claim a model's private memory or chat history is an authoritative project source.
  • Keep shared durable guidance in version-controlled, human-reviewable artifacts. Keep private project memory untracked unless sharing is explicitly authorized.
  • Let current source, Git state, tests, and runtime evidence override old notes.

Handoff check

Before finishing substantial work, make sure another contributor can identify:

  1. what is complete
  2. what is incomplete
  3. what evidence exists
  4. what must not be repeated or assumed
  5. the safest next action

Report the memory file used, stale material removed or flagged, related authorities updated, sensitive details excluded, and the next action.

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.