raia-live/amfs/cursor

Claude Code and agent skill by raia-live/amfs

Trust tier 2
  • community
LLM review pending
Source check pending

Checking that the SKILL.md file is still in the repository.

Guide agents on effective use of AMFS persistent memory — when to save, what to save, cost-conscious patterns, and session lifecycle. Use when agents interact with AMFS tools, persist knowledge, write memories, read briefings, or commit decision traces.

Install

npx vskill@latest install raia-live/amfs/cursor

SKILL.md

AMFS Agent Memory Guide

Persistent memory that survives across sessions, agents, and machines. This guide covers decision rules and conventions — not tool syntax (the MCP server already provides that).

Cost Model

OperationCostExamples
Read1 opamfs_read, amfs_search, amfs_list, amfs_recall, amfs_briefing, amfs_retrieve
Write2 opsamfs_write, amfs_record_context
CommitFREEamfs_commit_outcome — always commit, never skip

Session Lifecycle

Every session follows four phases:

  1. Identify — amfs_set_identity with a stable kebab-case role name (e.g. api-agent, infra-agent). Reuse the same name across conversations about the same domain. Don't use task-specific names like fix-button-color.
  1. Get Briefed — amfs_briefing returns compiled knowledge from the Cortex. One briefing (1 op) replaces many individual reads. Always start here.
  1. Work — read/write/record as you go. Write after meaningful work, not after every edit. Record decisions as they happen, not at the end.
  1. Commit — amfs_commit_outcome snapshots the full decision trace. This is free. Without it, the trace is lost when the session ends.

What to Save (worth 2 ops)

  • Decisions with rationale — why X over Y, trade-offs considered
  • Discovered patterns — reusable across sessions, link with pattern_refs
  • Risks and bugs — gotchas that would trip up future agents
  • Task summaries — what was done and key outcomes after significant work
  • External tool outputs — API results, metrics, incident data that informed decisions

What NOT to Save

  • Trivial edits — "renamed variable", "added comment" — VCS tracks these
  • File-level changes — save the decision, not the diff
  • Recomputable info — anything derivable from the codebase in under 5 seconds
  • Transient debug output — stack traces, test logs
  • One-sentence entries — too low signal for 2 ops; batch into richer entries

The test: "Would a future agent benefit from knowing this, or could they figure it out from the code?"

Cost-Conscious Patterns

  • Briefing first — amfs_briefing gives compiled context in 1 op vs. many individual reads
  • Batch writes — one rich entry beats three thin ones (2 ops vs. 6 ops)
  • Search before writing — avoid duplicating existing entries
  • Record context selectively — amfs_record_context costs 2 ops; use for meaningful decisions, not micro-steps
  • Always commit — amfs_commit_outcome is free and preserves the decision trace

Entity Paths

Use {repo}/{module} paths. Avoid generic paths like project or code.

myapp/auth, myapp/checkout, myapp/api, acme/infrastructure

Key Prefixes

PrefixUse case
pattern-Reusable patterns
risk-Known risks or bugs
decision-Architectural decisions
task-summary-What was done and why
action-Actions taken (experience log)

Memory Types

TypeUse whenDecay
fact (default)Stable knowledge — patterns, configs, verified decisionsNormal
beliefHypotheses, unverified observations (confidence < 0.9)2x faster
experienceActions taken, deployment logs1.5x slower

Confidence Scale

ScoreMeaning
1.0Verified fact, tested in production
0.7–0.9High confidence, not yet production-tested
0.4–0.6Hypothesis, needs validation
< 0.4Speculative, early signal

Beliefs should always be < 0.9. If you're sure enough for 0.9+, it's a fact.

Anti-Patterns

Don'tDo instead
Write every file editWrite the decision, not the change
Forget amfs_commit_outcomeAlways commit — it's free
Use generic entity pathsUse {repo}/{module}
Set confidence=1.0 on beliefsBeliefs are hypotheses — keep < 0.9
Skip amfs_briefingOne briefing replaces many reads
Write one-sentence entriesBatch into richer entries
Record every micro-stepRecord meaningful decisions only

Security

  • Security scanPassed, 100 of 100Pattern checks on the skill's files.
  • Model reviewNot runA model review is scheduled for the next rescan.

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

Repository stars
53
Repository forks
3
7-day trend score
20

Stars and forks are for the whole GitHub repository, not this skill alone. Not refreshed yet.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/raia-live/amfs/cursor/badge)](https://verified-skill.com/skills/raia-live/amfs/cursor)

Something wrong with this skill or its page? Tell us.