agentleFS
HenryQWHenryQW/skills/agent-memory/SKILL.md

agent-memory

Skill1 starsChanged 4 months agoApache-2.0View on GitHub

Use project-scoped agent memory with `.context/progress.md` and markdown files under `OBSIDIAN_PROJECT`. Use when the user or another workflow invokes `$agent-memory load`, `$agent-memory distill`, or `$agent-memory --setup`.

38 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use observability, data-analysis and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • OpenAISkill

    observability

    Vercel Observability expert guidance — Drains (logs, traces, speed insights, web analytics), Web Analytics, Speed Insights, runtime logs, custom events, OpenTelemetry integration, and monitoring dashboards. Use when instrumenting, debugging, or optimizing application performance and user experience on Vercel.

    Also in 4 projects that use this

  • shanggqmSkill

    data-analysis

    数据分析技能,提供结构化的数据查询、统计分析和报告生成工作流。当用户需要查询数据库、进行数据统计分析、生成分析报告时使用。按规范输出原始数据、计算结果、Markdown报告和HTML可视化报告。

    Also in 3 projects that use this

  • Microsoft AzureSkill

    skill-authoring

    Write Agent Skills that comply with the agentskills.io specification. WHEN: "create a skill", "new skill", "write a skill", "skill template", "skill structure", "review skill", "skill PR", "skill compliance", "SKILL.md format", "skill frontmatter", "skill best practices". DO NOT USE FOR: general documentation. INVOKES: vally CLI.

    Also in 3 projects that use this

  • VercelSkill

    technical-writing

    Write, edit, review, or audit user-facing documentation for the eve repository. Use for changes under docs/, documentation tied to eve APIs or CLI behavior, docs work based on Slack or support feedback, and requests to make eve docs clearer, more natural, or less AI-patterned while verifying claims against current source, tests, CLI help, public releases, and repository conventions.

    Also in 3 projects that use this

  • JetBrainsSkill

    webapp-testing

    Toolkit for interacting with and testing local web applications using Playwright. Supports verifying frontend functionality, debugging UI behavior, capturing browser screenshots, and viewing browser logs.

    Also in 4 projects that use this

  • AnthropicSkill

    pdf

    Comprehensive PDF manipulation toolkit for extracting text and tables, creating new PDFs, merging/splitting documents, and handling forms. When Claude needs to fill in a PDF form or programmatically process, generate, or analyze PDF documents at scale.

    Also in 5 projects that use this

The file

We split this skill into the rules it gives an agent

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

  • 0

    Common

    Most of the 33 edited versions we read keep these.

  • 1

    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:

    Only --setup reads references/setup.md or runs setup.

    line 14
  2. Distinctive:

    Read the 6,000-character-capped output only when notes loaded.

    line 48

“Common” is measured: kept by at least 6 in 10 of the 33 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

Agent Memory

Memory is artifact-driven: .context/progress.md is the task ledger; .context/decisions.jsonl stores durable candidates; generated load and preview artifacts remain under .context/. The project AGENTS.md declares OBSIDIAN_PROJECT; Agent/Memory/index.md links approved Decisions/ and Guidance/ notes. Confirmed notes are written there directly.

Only --setup reads references/setup.md or runs setup.

Boundary

  • $agent-memory load runs at a top-level workflow entry. Missing configuration or matching topics returns memory_load=SKIPPED and does not stop the caller.
  • During work, capture every distinct durable decision and guidance item; do not stop after the first candidate or combine unrelated topics.
  • A decision settles a non-obvious engineering choice. It qualifies when the user states it or approves a plan, design, or implementation containing it. Record its rationale, material alternatives or tradeoffs, and expected impact.
  • Guidance preserves a verified repository improvement future sessions should repeat or protect. Record what changed, when the practice applies, why it is better than the previous behavior, and what future sessions must watch. User corrections, repository invariants, and reusable failure lessons also qualify.
  • Capturing a candidate does not authorize final --apply.
  • Skip routine progress, bare file or commit summaries, checks, transient state, code-obvious mechanics, reversible local choices, and duplicates.
  • Memory should explain the project's engineering approach, not narrate task history.
  • $agent-memory distill runs only when the top-level workflow is otherwise ready to return final Done, Stop, or Blocked. Resumable PENDING, approval waits, handoffs, and nested skills preserve .context/decisions.jsonl without preview or apply.
  • Memory failure never replaces or hides the caller's terminal result.

Load

Select at most two exact linked topic slugs, then run:

python3 <agent_memory_dir>/scripts/memory_context.py --project-root /path/to/project --topic <topic-id> --out .context/memory-context.md

Read the 6,000-character-capped output only when notes loaded.

Capture and Distill

Append one decision candidate per choice:

python3 <agent_memory_dir>/scripts/append_decision.py --project-root /path/to/project --topic order-processing --decision "Keep retry policy in the application service" --reason "All transports share one policy." --alternatives "Per-adapter retries would diverge." --impact "Every transport now uses one policy." --source "approved architecture plan"

Append one guidance candidate per reusable practice:

python3 <agent_memory_dir>/scripts/append_decision.py --project-root /path/to/project --topic order-processing --guidance "Keep retry policy changes in the application service." --applies-when "Adding or changing an order transport." --change "Retry policy moved from adapters into the shared application service." --improvement "Transport behavior can no longer drift." --attention "Future sessions must keep transport adapters policy-free." --file src/orders/service.py --source "verified implementation"

Equivalent normalized candidates receive the same stable ID and are recorded once. Before the final preview, scan the accepted choices, repository delta, user corrections, reusable failures, and .context/progress.md for missed durable candidates. The ledger may confirm a candidate but is not itself memory. Then preview with:

python3 <agent_memory_dir>/scripts/distill_memory.py --project-root /path/to/project --source .context/decisions.jsonl

Inspect .context/memory-distill-preview.json; only explicit approval authorizes creating or updating its exact final notes and index links:

python3 <agent_memory_dir>/scripts/distill_memory.py --project-root /path/to/project --source .context/decisions.jsonl --apply

Apply rejects source or destination drift, writes directly to Decisions/ or Guidance/, and adds missing links to Agent/Memory/index.md. Do not apply before the user confirms the preview. Topic slugs map directly to final filenames; never search note bodies to choose a destination. Use .context/progress.md only to clarify an already durable decision. No durable records returns memory_write=SKIPPED without asking.

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.