Skip to content

ferdinandobons/knowledge-transfer / plugins/knowledge-transfer/skills/knowledge-transfer

knowledge-transfer

Verified

Use when a colleague is leaving a project and must hand over knowledge, or when someone new joins a project and receives handover.zip or a handover/ folder. Two phases: export (analyze project + agent project memories from Claude, Codex, or another assistant, then produce a transferable handover.zip archive containing a non-technical onboarding doc + filtered portable memories + a privacy-safe omissions manifest) and import (extract/read the handover package, dry-run the candidate memories, verify every claim against current code, then install accepted memories through the current agent's memory mechanism with per-memory receipts). Trigger on /knowledge-transfer, /knowledge-transfer:knowledge-transfer, $knowledge-transfer, 'handover', 'knowledge transfer', 'passaggio di consegne', 'onboard a new colleague', or when the user mentions taking over / leaving a project.

FileKindSize
SKILL.mdskill md11.9 KB
agents/openai.yamlconfig280 B
references/agent-memory.mdmarkdown2.5 KB
references/onboarding-template.mdmarkdown2.0 KB
references/package-format.mdmarkdown7.5 KB
references/privacy-filter.mdmarkdown3.0 KB

references/agent-memory.md

View on GitHub
# Agent memory handling

The handover package stores portable project memories as Markdown. The source and
destination memory stores are agent-specific, so do not assume every user is in
Claude Code.

## Export: find candidate memories

Use the memory paths and policies exposed by the current agent's system prompt or
developer instructions.

| Agent | Candidate source |
|---|---|
| Claude Code | The per-project memory directory for the current project, commonly under `~/.claude/projects/<project-slug>/memory/` when exposed by the environment. |
| Codex | The Codex memory registry under `$CODEX_HOME/memories` or `~/.codex/memories` when exposed. Search `MEMORY.md`/summaries for the current repo path, repo name, and project aliases, then open only the directly relevant memory or rollout files. |
| Other agents | The documented project memory store for that agent. If no path is visible, ask the user for the memory folder or proceed document-only. |

Scope strictly to the project being handed over. Exclude general user preferences,
unrelated projects, old global habits, and memory entries that cannot be tied back
to this repo.

## Import: install accepted memories

Respect the current agent's memory policy. The user invoking import is asking to
install project knowledge, but the agent may still require a specific update path.

| Agent | Installation behavior |
|---|---|
| Claude Code | Write accepted memories into the current project's Claude memory directory and update its `MEMORY.md` index if present. |
| Codex | Do not edit generated memory indexes directly unless the current instructions allow it. Prefer the approved Codex memory-update mechanism, such as an ad-hoc note under `$CODEX_HOME/memories/extensions/ad_hoc/notes/`, containing the accepted portable memories and index summaries. |
| Other agents | Convert the accepted memory bodies to that agent's documented memory format, or present an importable bundle if no writable memory store is available. |

Never bypass a policy that says memory writes are restricted. If direct installation
is blocked, report the accepted memories and the exact reason installation was not
performed.

Before writing, always complete the dry-run import plan described in
`package-format.md`. After each candidate memory is handled, write a receipt whose
`destination` field names the actual storage path, approved memory-update
mechanism, or blocked reason. Examples:

- `claude-project-memory:<path>`
- `codex-ad-hoc-memory-note:<path>`
- `presented-only:memory-policy-blocked`
- `rejected:stale-no-evidence`
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 470kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall