ferdinandobons/knowledge-transfer / plugins/knowledge-transfer/skills/knowledge-transfer
knowledge-transfer
VerifiedUse 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.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 11.9 KB |
| agents/openai.yaml | config | 280 B |
| references/agent-memory.md | markdown | 2.5 KB |
| references/onboarding-template.md | markdown | 2.0 KB |
| references/package-format.md | markdown | 7.5 KB |
| references/privacy-filter.md | markdown | 3.0 KB |
references/package-format.md
View on GitHub# Handover package format
The export phase writes a local staging folder and a transfer archive into the
root of the target project:
```
handover.zip # transfer artifact to send to the next colleague
handover/
ONBOARDING.md # onboarding document (see onboarding-template.md)
memories/
<slug>.md # one file per exported memory
omissions.json # privacy-safe omitted-memory accounting
manifest.json # export metadata
```
`handover/` is a working folder created where the export runs. `handover.zip` is
the artifact to pass to the next person. Do not commit either file by default;
only commit them if the team explicitly decides to store handovers in the repo.
The zip archive must contain the top-level `handover/` directory, not just its
contents:
```
handover.zip
└── handover/
├── ONBOARDING.md
├── manifest.json
├── omissions.json
└── memories/
└── <slug>.md
```
The zip wrapper is separate from the package schema. The folder format inside
the archive is versioned by `manifest.version`. New exports use version `2`.
Importers should support version `1` as a legacy format when possible.
## manifest.json
```json
{
"version": 2,
"exported_at": "2026-06-10",
"author": "Jane Doe",
"commit": "a1b2c3d4e5f6a7b8c9d0e1f2a3b4c5d6e7f8a9b0",
"language": "en",
"memories": {
"exported": 3,
"excluded": 1
},
"artifacts": {
"omissions": "omissions.json"
}
}
```
Field rules:
- `version` — `2` for new exports. Importers must refuse unknown versions. Version
`1` packages may omit `omissions.json`, source hashes, and claim metadata.
- `exported_at` — ISO date of the export.
- `author` — from `git config user.name`. Omit the field if unavailable. (The privacy
filter applies to memory *content*, not to git metadata already public in the repo.)
- `commit` — full SHA of `HEAD` at export time. Omit if the project is not a git repo.
The import phase uses it to diff what changed since the export.
- `language` — ISO 639-1 code of the language ONBOARDING.md is written in.
- `memories.exported` / `memories.excluded` — counts after the privacy filter. They let
the importer (and the team) see that filtering happened.
- `artifacts.omissions` — relative path to the privacy-safe omissions manifest.
## omissions.json
Privacy-safe accounting for memories that were excluded or materially redacted.
It must never include the omitted private content itself.
```json
{
"version": 1,
"policy": "aggregate categories only; no omitted private content",
"total_omitted": 2,
"categories": [
{
"category": "personal_profile",
"count": 1,
"reason": "User-specific preference or identity data is not project knowledge."
},
{
"category": "secret",
"count": 1,
"reason": "Secrets are excluded entirely, not redacted into the package."
}
]
}
```
Allowed categories:
- `personal_profile` — outgoing colleague identity, preferences, role, or personal
operating style.
- `secret` — token, password, key, credential, or secret-bearing URL.
- `private_link` — personal dashboard, private gist, private document, or
non-team-accessible pointer.
- `machine_path` — absolute local path outside the repo that cannot be made
repo-relative.
- `not_project_knowledge` — unrelated project memory or generic user habit.
- `other_private` — use sparingly when a safer category does not fit.
Safe omission rules:
- Include counts and generic categories only.
- Do not include original text, names, emails, usernames, domains that expose a
private system, token prefixes, or file paths outside the repo.
- If a one-line reason would reveal the omitted content, use the category's
generic reason instead.
## memories/<slug>.md
Portable Markdown memory format. It intentionally matches the simple frontmatter
shape used by Claude Code memories and is easy for Codex or another agent to
translate into its own memory-update mechanism:
```markdown
---
name: <slug>
description: <one-line summary used for relevance during recall>
metadata:
type: project | reference | feedback
source_id: sha256:<12 hex chars>
source_hash: sha256:<64 hex chars>
redactions:
count: 2
categories:
- personal_identifier
- machine_path
claims:
- "Repo-relative file path src/api/client.ts defines the API client used by the dashboard."
- "npm run test is the project test command documented in package.json."
---
<body>
```
Constraints:
- Type `user` NEVER appears in a package (see privacy-filter.md).
- Bodies are rewritten in neutral, impersonal voice — no names, no personal data.
- `source_id` and `source_hash` are derived from the privacy-normalized source
memory before final rewrite. Private values must be removed or replaced before
hashing. Include hashes only for exported memories, never for omitted private
or secret-bearing memories.
- `redactions.categories` uses safe category names only, never private values.
- `claims` is a short list of concrete facts the import phase can verify against
the current repo.
- Every memory must be verifiable: if it cites files, paths, or identifiers, they must
exist in the project at export time.
- `[[name]]` links between memories are allowed only if the target memory is also
exported; otherwise remove the link syntax and keep plain text.
- Importers may adapt the storage wrapper or index to Claude, Codex, or another
agent, but must preserve the verified project fact in the memory body.
## Import artifacts
Import creates local audit files inside the extracted `handover/` folder. They are
not present in the original export archive.
```
handover/
import-plan.json
import-report.json
import-receipts/
<slug>.json
```
### import-plan.json
Written before any memory store is modified. This is the dry-run output.
```json
{
"version": 1,
"generated_at": "2026-06-10T12:00:00Z",
"mode": "dry-run",
"package_commit": "a1b2c3d4e5f6a7b8c9d0e1f2a3b4c5d6e7f8a9b0",
"candidates": [
{
"memory_file": "memories/api-client.md",
"source_id": "sha256:7f83b1657ff1",
"source_hash": "sha256:7f83b1657ff1fc53b92dc18148a1d65dfc2d4b1fa3d677284addd200126d9069",
"claims": ["src/api/client.ts defines the API client."],
"evidence_checked": ["exists: src/api/client.ts"],
"proposed_action": "install",
"reason": "All claims verified against the current repo.",
"destination": "codex-ad-hoc-memory-note"
}
]
}
```
### import-receipts/<slug>.json
One receipt per candidate memory, including accepted, rewritten, rejected, and
blocked memories.
```json
{
"version": 1,
"memory_file": "memories/api-client.md",
"source_id": "sha256:7f83b1657ff1",
"source_hash": "sha256:7f83b1657ff1fc53b92dc18148a1d65dfc2d4b1fa3d677284addd200126d9069",
"export_commit": "a1b2c3d4e5f6a7b8c9d0e1f2a3b4c5d6e7f8a9b0",
"claims": ["src/api/client.ts defines the API client."],
"evidence_checked": ["exists: src/api/client.ts"],
"stale_or_superseded_findings": [],
"privacy_redactions": {
"count": 0,
"categories": []
},
"destination": "codex-ad-hoc-memory-note",
"final_action": "installed",
"reason": "All claims verified and memory was written through the approved memory-update mechanism."
}
```
`final_action` must be one of:
- `installed`
- `rewritten`
- `rejected`
- `blocked`
### import-report.json
Summary file with counts and receipt paths:
```json
{
"version": 1,
"installed": 2,
"rewritten": 1,
"rejected": 1,
"blocked": 0,
"receipts": [
"import-receipts/api-client.json"
]
}
```
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| 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: 471k | Updated | Claude CodeCodex | Verified | Install |
| 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: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| 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 available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |