/contrib-consolidate
Internal phase — consolidate a contributor's L1 atoms into an L3 persona (11 dimensions). Invoked by contrib-profile / `/contrib build`.
$ npx -y skills add baodq97/tencentdb-agent-memory --skill contrib-consolidate --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
- Fires itselfAuto-invocation. Claude auto-loads it when your prompt matches the work.
- You can call itInvoke it directly when you want it.
- Slash command
/contrib-consolidate
Context preview
The summary Claude sees to decide when to auto-load this skill.
Internal phase — consolidate a contributor's L1 atoms into an L3 persona (11 dimensions). Invoked by contrib-profile / `/contrib build`.
SKILL.md
contrib-consolidate.SKILL.mdname: contrib-consolidate
description: Internal phase — consolidate a contributor's L1 atoms into an L3 persona (11 dimensions). Invoked by contrib-profile / `/contrib build`.
user-invocable: false
Contributor Consolidate
Group one subject's L1 atoms into themes (L2 scenes, conceptual) and write a single L3 persona summarising all 11 dimensions with evidence.
Workflow
1. Read the atoms
tmem contrib atoms <subject-id>
2. Build the persona
**Read `references/persona-guide.md` first** — it covers merging atoms (not listing), resolving conflicting atoms, weighting by evidence, when to mark "insufficient data", and the quality bar (a persona should predict how the subject tackles a *new* task).
For each of the 11 dimensions (`idea, plan, solve, craft, comms, mentor, conflict, scope, ownership, execution`), synthesise the atoms in that dimension into 1–3 sentences. Carry the strongest evidence links into the text. If a dimension has no atoms, set it to `"insufficient data"`.
Collect 3–6 `notable_traits` — distinctive things that don't fit a fixed dimension (e.g. "writes prose-quality commit bodies", "prefers small composable modules").
3. Write the persona
tmem contrib upsert-persona --json '{
"subject_id": "<id>",
"summary": "<2-3 sentence overview of how this engineer works>",
"dimensions": {
"idea": "...", "plan": "...", "solve": "...", "craft": "...",
"comms": "...", "mentor": "...", "conflict": "...",
"scope": "...", "ownership": "...", "execution": "..."
},
"notable_traits": ["...", "..."],
"updated_time": "<ISO timestamp>"
}'4. Report
Print the persona (`tmem contrib persona <id>`) and tell the user which dimensions are well-evidenced vs "insufficient data".
Read more
name: contrib-consolidate description: Internal phase — consolidate a contributor's L1 atoms into an L3 persona (11 dimensions). Invoked by contrib-profile / `/contrib build`. user-invocable: false
Contributor Consolidate
Group one subject's L1 atoms into themes (L2 scenes, conceptual) and write a single L3 persona summarising all 11 dimensions with evidence.
Workflow
1. Read the atoms
tmem contrib atoms <subject-id>
2. Build the persona
**Read `references/persona-guide.md` first** — it covers merging atoms (not listing), resolving conflicting atoms, weighting by evidence, when to mark "insufficient data", and the quality bar (a persona should predict how the subject tackles a *new* task).
For each of the 11 dimensions (`idea, plan, solve, craft, comms, mentor, conflict, scope, ownership, execution`), synthesise the atoms in that dimension into 1–3 sentences. Carry the strongest evidence links into the text. If a dimension has no atoms, set it to `"insufficient data"`.
Collect 3–6 `notable_traits` — distinctive things that don't fit a fixed dimension (e.g. "writes prose-quality commit bodies", "prefers small composable modules").
3. Write the persona
tmem contrib upsert-persona --json '{
"subject_id": "<id>",
"summary": "<2-3 sentence overview of how this engineer works>",
"dimensions": {
"idea": "...", "plan": "...", "solve": "...", "craft": "...",
"comms": "...", "mentor": "...", "conflict": "...",
"scope": "...", "ownership": "...", "execution": "..."
},
"notable_traits": ["...", "..."],
"updated_time": "<ISO timestamp>"
}'4. Report
Print the persona (`tmem contrib persona <id>`) and tell the user which dimensions are well-evidenced vs "insufficient data".
Four-layer long-term memory (L0 Conversation → L1 Atom → L2 Scene → L3 Persona) for Claude Code, inspired by Tencent/TencentDB-Agent-Memory. Fully local — no external Gateway, no paid API, no Python.
Repo: baodq97/tencentdb-agent-memory
Other skills on tencentdb-agent-memory.
- /contrib-ingest
Internal phase — extract L1 contributor atoms from a GitHub subject's raw activity. Invoked by contrib-profile / `/contrib ingest`.
Open skill - /contrib-profile
Orchestrator for Contributor Intelligence. Trigger when the user pastes a GitHub profile/repo URL or handle and asks to profile, analyze, learn from, or study an engineer (EN/VI — "phân tích người này", "học từ người này"). Runs add→ingest→build→playbook end-to-end via the
Open skill - /contrib-synthesize
Internal phase — synthesize learnable playbooks, cross-engineer common capabilities, and you-vs-role-model comparisons from built personas. Invoked by contrib-profile / `/contrib playbook|compare`.
Open skill - /memory-consolidate
Consolidate L1 memory atoms into L2 scene blocks and L3 persona. Invoked by the memory-consolidator agent, or manually via /memory-consolidate.
Open skill - /memory-seed
Extract L1 memory atoms from Claude Code conversation history. Run manually via /memory-seed.
Open skill - /memory-view
Open the memory visualiser in the browser — a health check on the memory store: is it working, what is wrong, and what to run to fix it. Run manually via /memory-view.
Open skill

