sandrinio/wikeeper
A Karpathy-style LLM wiki that your AI coding agent keeps. Gated ingest, cited facts, search. Claude Code · Codex · Antigravity.
ARCHETYPE
Orchestrator
A bench of specialised subagents. The main agent mostly delegates.
CONTEXT TAX · EVERY TURN
up to ~6.7k tokens
Claude Code loads MCP tools on demand, so the real cost is usually lower.
Heavy · median rig: 2.3k · breakdown
Copy this rig
# review before running: this installs third-party code $ claude mcp add cleargate -- npx -y cleargate@0.27.0 mcp serve $ npx degit sandrinio/wikeeper/.claude ./rig-wikeeper # inspect, then merge into .claude/
MCP servers are added to Claude Code at local scope; env vars are shown as YOUR_… placeholders — we never store values. Files are fetched with degit into a separate folder so you can review before merging.
$ claude mcp add cleargate -- npx -y cleargate@0.27.0 mcp serve $ npx degit sandrinio/wikeeper/.claude/skills/flashcard .claude/skills/flashcard $ npx degit sandrinio/wikeeper/.claude/skills/sprint-dashboard .claude/skills/sprint-dashboard $ npx degit sandrinio/wikeeper/.claude/skills/sprint-execution .claude/skills/sprint-execution $ npx degit sandrinio/wikeeper/templates/skills/wiki-ingest .claude/skills/wiki-ingest $ npx degit sandrinio/wikeeper/templates/skills/wiki-search .claude/skills/wiki-search $ npx degit sandrinio/wikeeper/templates/skills/wiki-setup .claude/skills/wiki-setup
$ curl -fsSL --create-dirs -o .claude/agents/architect-reader.md https://raw.githubusercontent.com/sandrinio/wikeeper/main/.claude/agents/architect-reader.md $ curl -fsSL --create-dirs -o .claude/agents/architect-synth.md https://raw.githubusercontent.com/sandrinio/wikeeper/main/.claude/agents/architect-synth.md $ curl -fsSL --create-dirs -o .claude/agents/architect.md https://raw.githubusercontent.com/sandrinio/wikeeper/main/.claude/agents/architect.md $ curl -fsSL --create-dirs -o .claude/agents/cleargate-wiki-contradict.md https://raw.githubusercontent.com/sandrinio/wikeeper/main/.claude/agents/cleargate-wiki-contradict.md $ curl -fsSL --create-dirs -o .claude/agents/cleargate-wiki-ingest.md https://raw.githubusercontent.com/sandrinio/wikeeper/main/.claude/agents/cleargate-wiki-ingest.md $ curl -fsSL --create-dirs -o .claude/agents/cleargate-wiki-lint.md https://raw.githubusercontent.com/sandrinio/wikeeper/main/.claude/agents/cleargate-wiki-lint.md $ curl -fsSL --create-dirs -o .claude/agents/cleargate-wiki-query.md https://raw.githubusercontent.com/sandrinio/wikeeper/main/.claude/agents/cleargate-wiki-query.md $ curl -fsSL --create-dirs -o .claude/agents/developer.md https://raw.githubusercontent.com/sandrinio/wikeeper/main/.claude/agents/developer.md $ curl -fsSL --create-dirs -o .claude/agents/devops.md https://raw.githubusercontent.com/sandrinio/wikeeper/main/.claude/agents/devops.md $ curl -fsSL --create-dirs -o .claude/agents/qa.md https://raw.githubusercontent.com/sandrinio/wikeeper/main/.claude/agents/qa.md $ curl -fsSL --create-dirs -o .claude/agents/reporter.md https://raw.githubusercontent.com/sandrinio/wikeeper/main/.claude/agents/reporter.md
Merge into .claude/settings.json (project) or ~/.claude/settings.json (user). Hook commands reference scripts in the source repo — copy those too.
{
"hooks": {
"PreToolUse": [
{
"matcher": "Task|Agent",
"hooks": [
{
"type": "command",
"command": "${CLAUDE_PROJECT_DIR}/.claude/hooks/pending-task-sentinel.sh --host claude"
},
{
"type": "command",
"command": "${CLAUDE_PROJECT_DIR}/.claude/hooks/pre-tool-use-task.sh"
}
]
},
{
"matcher": "Edit|Write",
"hooks": [
{
"type": "command",
"command": "${CLAUDE_PROJECT_DIR}/.claude/hooks/pre-edit-gate.sh --host claude"
}
]
},
{
"matcher": "AskUserQuestion",
"hooks": [
{
"type": "command",
"command": "${CLAUDE_PROJECT_DIR}/.claude/hooks/pre-tool-use-autonomy.sh"
}
]
}
]
}
} MCP servers (1)
| server | source | est. tokens |
|---|---|---|
| cleargate | npm | 2.5k |
Skills (6)
Subagents (11)
| architect-reader model: haiku | Use DURING Sprint Design Review fan-out (STORY-033-03 / EPIC-033). One instance spawned per story by the Orchestrator. Reads a single story file and returns a structured digest for architect-synth to |
| architect-synth model: opus | Use DURING Sprint Design Review fan-out (STORY-033-03 / EPIC-033) AFTER all architect-reader agents have returned digests. Consumes N reader digests, evaluates the real-collision predicate, forms cont |
| architect model: opus | Use BEFORE development starts on a ClearGate sprint milestone. Reads the story file + relevant existing code, produces a tight implementation sketch (files to touch, schema deltas, test shape, risks) |
| cleargate-wiki-contradict model: sonnet | Use during ingest Phase 4 to perform a neighborhood-scoped contradiction check on a draft or in-review wiki page. Advisory only — always exits 0. Emits zero or more `contradiction:` finding lines plus |
| cleargate-wiki-ingest model: haiku | Use AFTER any Write or Edit on a raw work item under .cleargate/delivery/**. Reads the just-written file, creates or updates one wiki page at .cleargate/wiki/{epics|stories|sprints|proposals|crs|bugs} |
| cleargate-wiki-lint model: sonnet | Use BEFORE Gate 1 (Proposal approval) and Gate 3 (Push) to enforce wiki-vs-raw consistency. Default mode (enforcement) exits non-zero on any drift, naming the offending page; halts gate transitions. ` |
| cleargate-wiki-query model: haiku | Use AT TRIAGE before drafting any new Proposal, Epic, or Story — surfaces prior work that may already cover the topic. Read-only by default. Reads .cleargate/wiki/index.md first, then targeted per-ite |
| developer model: sonnet | Use to implement one ClearGate Story end-to-end. Reads the story file + the Architect's plan, writes production code + unit tests in the designated worktree, runs typecheck and tests locally, commits |
| devops model: sonnet | Use AFTER QA-Verify pass + Architect post-flight pass on a Story. Owns mechanical merge, worktree teardown, state transition to Done, mirror parity diff post-merge. Never authors code. Halts on confli |
| qa model: sonnet | Use AFTER a Developer agent reports STATUS=done on a Story. Independent verification gate. Re-runs typecheck + tests in a fresh shell, diffs the commit against the Story's acceptance Gherkin, flags mi |
| reporter model: sonnet | Use ONCE at the end of a ClearGate sprint, after all stories have passed QA. Synthesizes the token ledger, flashcards, git log, DoD checklist, and story files into a sprint report using the Sprint Rep |
Hooks (7)
| event | matcher | runs |
|---|---|---|
| SubagentStop | * | ${CLAUDE_PROJECT_DIR}/.claude/hooks/token-ledger.sh |
| PreToolUse | Task|Agent | ${CLAUDE_PROJECT_DIR}/.claude/hooks/pending-task-sentinel.sh --host claude |
| PreToolUse | Task|Agent | ${CLAUDE_PROJECT_DIR}/.claude/hooks/pre-tool-use-task.sh |
| PreToolUse | Edit|Write | ${CLAUDE_PROJECT_DIR}/.claude/hooks/pre-edit-gate.sh --host claude |
| PreToolUse | AskUserQuestion | ${CLAUDE_PROJECT_DIR}/.claude/hooks/pre-tool-use-autonomy.sh |
| PostToolUse | Edit|Write | ${CLAUDE_PROJECT_DIR}/.claude/hooks/stamp-and-gate.sh |
| SessionStart | * | ${CLAUDE_PROJECT_DIR}/.claude/hooks/session-start.sh |
Similar rigs
vannt-dev/governed-agent-sdlc
Ship with agents. Keep humans in control. Cross-platform governance toolkit for AI coding agents.
Fort Knox 632 tok · · Q24
yeyo11/agentry
Run Claude Code as a service: REST API, web UI and multi-agent orchestration in one container
Pragmatist 5.2k tok · · Q18
ThinhHoang1/claude-code-agent-team
An AI dev team for Claude Code agent teams: 8 role agents, a HANDOFF protocol and 3 hook quality gates — with the TeammateIdle infinite-loop bug fixed and regression-tested.
Orchestrator 1.6k tok · · Q15
adyusuf/claude-code-standards
The live Claude Code rule set I ship production software with — 33 rules, 22 standards, 14 agent roles, each traced to the measurement or incident that produced it.
Orchestrator 6.9k tok · · Q10