starmynd-org/infinite-brain-os
Agent Claude Code
Personal-repo curation agent. Sweeps this personal repo's scratch-lifecycle namespaces and proposes which ones to promote, merge, or archive. Lighter than the canon variant: this curator works on a single repo.
starmynd-org/infinite-brain-os
Agent Claude Code
The agent that proposes canon writes and revisions for a namespace. It compresses operator-validated synthesis into core-doctrine, carries derivedfrom provenance, verifiedat and verifiedby, and a changelog, and stops at the approval gate. The operator is the only approver of canon; this agent drafts, it never promotes.
starmynd-org/infinite-brain-os
Agent Claude Code
The agent that turns raw archive and provenance into derived thinking. It writes within-namespace synthesis into knowledge/ /synthesis/ and cross-namespace synthesis into the root synthesis/, detects where nodes and sources disagree, and packages canon-candidates for operator review. Synthesis is interpretive and…
starmynd-org/infinite-brain-os
Agent Claude Code
The agent that reviews knowledge freshness scoped by posture. It checks review-on-edit, periodic, and live namespaces differently, flags nodes whose verifiedat is stale relative to their posture, and proposes re-verification. It never puts a periodic timer on stable doctrine and never silently rewrites a stale claim.
starmynd-org/infinite-brain-os
Agent Claude Code
A specialist architect agent for deciding what entity mix a problem needs, what should remain human-gated, and which builder skills to apply next.
starmynd-org/infinite-brain-os
Agent Claude Code
The agent that processes the durable intake layer. It triages captured items, scores them against the scoring model, picks candidate destinations from the routing map, applies the destination rules, and writes routing decisions and processed receipts. It respects the no-live-queue-in-git boundary: live queues stay in…
starmynd-org/infinite-brain-os
Agent Claude Code
A specialist agent for shaping autoresearch-style and standing-runtime feedback loops into explicit Infinite Brain artifacts without inventing a new entity.
starmynd-org/infinite-brain-os
Agent Claude Code
Per-namespace maintenance agent. Given one namespace, it sweeps the INDEX retrieval router, canon health, structural and profile lint, and freshness, then returns a prioritized maintenance report with proposed edits. It surfaces and recommends; the operator approves canon changes.
starmynd-org/infinite-brain-os
Agent Claude Code
The lint agent for a namespace. It runs bash system/validate.sh for the deterministic checks, adds the profile-aware fuzzy review that validate.sh cannot do, and returns a findings report split into deterministic errors, deterministic warnings, and fuzzy observations. It reports; it does not auto-fix canon.
starmynd-org/infinite-brain-os
Agent Claude Code
A subagent that retrieves relevant knowledge nodes by reading the local git tree, supplements with a web search, and produces a structured research summary with citations and suggested follow-up nodes.
starmynd-org/infinite-brain-os
Agent Claude Code
Keeps the root sessions ledger clean, complete, and promotive rather than letting useful chat history die in logs.
starmynd-org/infinite-brain-os
Agent Claude Code
Triage inbound customer emails for Emberline Candle Studio: classify each message, draft replies for routine ones in the brand voice, and escalate refunds and complaints for human review.
starmynd-org/infinite-brain-os
Command Claude Code
Use when an operator wants to create or refine a department, its head agent, intake posture, linked tools and workflows, and its role in the AI shadow department model.
starmynd-org/infinite-brain-os
Command Claude Code
Use when an operator wants to create a new real knowledge namespace or migrate an existing corpus into Infinite Brain V2 shape, using normal English instead of rigid flags.
starmynd-org/infinite-brain-os
Command Claude Code
Close a tracked AI session: write the closeout review, extract memory and task and swarm candidates, and finalize the session record.
starmynd-org/infinite-brain-os
Command Claude Code
Slash command that creates a new namespace file under system/namespaces/ in the personal repo, with lifecyclestate: scratch.
starmynd-org/infinite-brain-os
Command Claude Code
Use when a person wants the brain to learn their business and tell them what to build: runs the full interview, recommendation, and scoping loop. Also usable mid-life to diff an evolved business against the existing architecture.
starmynd-org/infinite-brain-os
Command Claude Code
Promote a node from this personal repo to a department or company-canon repo. Rewrites frontmatter, resolves wikilinks, copies the file across repos, and opens a PR against the target repo.
starmynd-org/infinite-brain-os
Command Claude Code
Start a tracked AI session: create the session record, declare the transcript path, and record the initial context and linked work.
starmynd-org/infinite-brain-os
Command Claude Code
Produce the Emberline morning brief: open orders, low-stock candles, unanswered customer emails, and one suggested focus for the day, written to outputs/.
starmynd-org/infinite-brain-os
Hook Claude Code
Runs when a session starts, executing session-start.sh via bash. From starmynd-org/infinite-brain-os.
starmynd-org/infinite-brain-os
Hook Claude Code
Runs after a tool call finishes for Write and Edit tool calls, executing post-write-lint.sh via bash. From starmynd-org/infinite-brain-os.
starmynd-org/infinite-brain-os
Hook Claude Code
Runs when the agent finishes a response, executing stop-check.sh via bash. From starmynd-org/infinite-brain-os.
starmynd-org/infinite-brain-os
Settings file Claude Code
Agent settings declaring 3 hook events (SessionStart, PostToolUse, Stop).