Engineering AgentMemoryLearning

Self-Improving — Remember

Save important patterns to auto-memory for future sessions.

$skillfish add ~skills/self-improving-agent-remember

Real Examples Get Started Models & Context

Real Conversations

See how the Self-Improving — Remember skill responds to actual requests. These are representative examples of what your AI agent will produce.

You

Just spent 2 hours debugging a CORS issue. Turns out our CDN (Cloudflare) strips the Authorization header before it reaches the backend — it's not a backend config problem. Save this so I never debug it from scratch again.

Agent

That's exactly the kind of hard-won insight that deserves explicit memory — non-obvious, non-reproducible without the original context.

Checking for duplicates... No existing entry about CORS or Cloudflare Authorization header stripping.

Writing to MEMORY.md:

- CORS: Cloudflare strips `Authorization` header before requests reach backend.
  Not a backend config issue — check CDN transform rules first.

Saved to auto-memory

  "CORS: Cloudflare strips Authorization header —
   check CDN rules before backend config"

  MEMORY.md: 142/200 lines
  Claude will see this at the start of every session in this project.

Suggestion: This sounds like a CLAUDE.md debugging rule — "When investigating CORS errors, check Cloudflare transform rules before backend config." A rule makes Claude apply this heuristic proactively, not just recall it as background context.

Want to /si:promote it to a debugging rule, or leave as a memory note?

You

Remember that our integration tests need TEST_DB_URL pointing to the test database and you have to run pnpm db:seed:test first or everything fails. The seed takes about 8 seconds.

Agent

Checking for duplicates... Found: "Integration tests require TEST_DB_URL" (line 89) — similar but missing the seed step and timing detail. Updating that entry.

Updated MEMORY.md line 89:

- Integration tests: set `TEST_DB_URL=postgres://localhost:5432/test_db` and
  run `pnpm db:seed:test` (~8s) before `pnpm test:integration` or all tests fail

Updated auto-memory (existing entry enriched, not duplicated)

  "Integration tests: TEST_DB_URL + pnpm db:seed:test (~8s) before running suite"

  MEMORY.md: 143/200 lines
  Claude will follow this sequence when asked to run integration tests.

This is a CLAUDE.md rule candidate — it's a required sequence, not discoverable context. A rule makes Claude run the seed step automatically without you reminding it.

Draft rule:

## Testing
- Integration tests: `TEST_DB_URL` must be set + `pnpm db:seed:test` (~8s) must run first

Promote now, or let it prove itself over a few more sessions?

Get Started

How to install and use this skill in your preferred environment.

Skills are designed for AI coding agents (Claude Code, Cursor, Windsurf) and IDE-based workflows where the agent can read files, run scripts, and act on your codebase. Web-based AI can use the knowledge and frameworks, but won't have tool access.

Models & Context

Which AI models and context windows work best with this skill.

Recommended Models

Best

Claude Opus 4 Claude Sonnet 4 GPT-4.1 Gemini 2.5 Pro Grok 3 Kimi K2

Good

Claude Haiku 4.5 GPT-4.1 mini Gemini 2.5 Flash Grok 3 mini

Larger models produce more detailed, production-ready outputs.

Context Window

This skill's SKILL.md is typically 3–10 KB — fits in any modern context window.

8K Skill only

32K+ Skill + conversation

100K+ Skill + references + codebase

All current frontier models (Claude, GPT, Gemini) support 100K+ context. Use the full window for complex multi-service work.

Pro tips for best results

1

Be specific

Include numbers — users, budget, RPS — so the skill can size the architecture.

2

Share constraints

Compliance needs, team size, and existing stack all improve the output.

3

Iterate

Start with a high-level design, then ask follow-ups for IaC, cost analysis, or security review.

4

Combine skills

Pair with companion skills below for end-to-end coverage.

Ready to try Self-Improving — Remember?

Install the skill and start getting expert-level guidance in your workflow — any agent, any IDE.

$skillfish add ~skills/self-improving-agent-remember

← Browse all 169 skills