Skip to content
Agentmart
Submit a tool

ai-orchestration-learning

Use at the end of a task in Claude Code when the work produced verified, reusable project knowledge worth saving with the MCP memory.learn tool.

Skill by ufukyuksel-dev

IndexedMIT
Write a review

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/ufukyuksel-dev-ai-orchestration-learning

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Use at the end of a task in Claude Code when the work produced verified, reusable project knowledge worth saving with the MCP memory.learn tool.

Use cases: productivity, agent-building

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/ai-orchestration-learning

What's inside

Skill name
ai-orchestration-learning
Description
Use at the end of a task in Claude Code when the work produced verified, reusable project knowledge worth saving with the MCP memory.learn tool.
License
MIT
Instructions
38 lines in SKILL.md · view on GitHub

Files (1)

  • SKILL.md

SKILL.md (first lines)

# Project learning — Claude Code

Call `memory.learn(rootPath, candidates)` **at most once**, at task end, only when the task established knowledge that a future session would otherwise have to rediscover. No durable finding means no call. If `ai-orchestration` is unavailable or a call failed with a connection error, do not retry; just finish the task.

## Worth saving
Verified, project-specific, reusable, non-obvious:
- a decision and why (`decision`), a team convention (`convention`), a multi-step procedure (`procedure`)
- behaviour that is surprising from the code alone (`behavior`), a known failure cause (`debug_finding`)
- where a change must be made (`change_point`), how to test something (`test_address`), a data/field mapping (`field_mapping`)

## Never save
Progress notes, changed-file lists, test-pass reports, facts obvious from one read of the code, guesses, secrets, rules (rules use `rules.draft`), or anything `memory.search` already returned.

## Candidate shape
`{kind, summary (≤160 chars), content (≤700 chars, one fact), locators, appliesWhen?, limitations?, reusableFor?}`, 1–3 per call, at most 8 distinct locators in total.

- Code knowledge (`behavior`, `navigation`, `field_mapping`, `debug_finding`, `change_point`, `test_address`) needs at least one symbol or file locator copied exactly from the source you read:
  - class: `{"kind":"symbol","ref":"com.example.PaymentService","path":"src/main/java/com/example/PaymentService.java"}`
  - method: `{"kind":"symbol","ref":"com.example.PaymentService#approve","signature":"approve(java.lang.String)","path":"src/main/java/com/example/PaymentService.java","role":"primary_change_point"}`
  - file: `{"kind":"file","ref":"src/main/resources/application.yml","role":"configuration"}`
  - A directory locator is navigation only; it cannot be the sole locator.
- `decision`, `convention` and `procedure` may use `"locators": []`.
- Never invent scanner or database IDs.

## Examples
Good:
1. `{"kind":"debug_finding","summary":"Refund approval fails when the ledger lock is taken twice","content":"PaymentService.approve acquires the ledger lock and RefundJob.run acquires it again; running both in one transaction deadlocks. Refunds must go through RefundJob only.","locators":[{"kind":"symbol","ref":"com.example.PaymentService#approve","path":"src/main/java/com/example/PaymentService.java"}]}`
2. `{"kind":"decision","summary":"Invoices keep two decimal places because of the tax export","content":"Invoice amounts are rounded HALF_EVEN to two decimals because the national tax export rejects more precision (finance decision FIN-12).","locators":[]}`
3. `{"kind":"procedure","summary":"Run integration tests offline","content":"Use ./mvnw -q -o verify -Pit; the online build cannot reach the artifact proxy from CI runners.","locators":[]}`

Bad (do not send):
1. "Fixed the failing test by changing >= to ==" — a progress note.
2. "PaymentService is a Spring service" — obvious from the code.
3. "The timeout is probably configurable" — a guess, not verified.

The response reports `created`, `reused` (an existing learned record matched), `duplicates` (the gate refused an already-covered fact; nothing saved) and `rejected`. Reused and duplicate are success, not errors — do not resend. Report only what the tool returned.

From ufukyuksel-dev/ai-orchestration, MIT.

More from ufukyuksel-dev/ai-orchestration

Signals

Stars
0
Forks
0
Open issues
0
Last commit (this folder)
Sep 27, 2026 (1 month ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from ufukyuksel-dev

Cite this page

ai-orchestration-learning skill. Agent Mart. https://agentmart.dev/skills/ufukyuksel-dev-ai-orchestration-learning (last updated 2026-10-24)