Skill Claude CodeCodex
Web research engine that acquires knowledge on a named topic from external sources (WebSearch + WebFetch across multiple queries and depths) and writes the findings as structured knowledge-tree nodes with citations. Use whenever the user says "research X", "learn about Y", "build a knowledge base on Z", or the…
Skill Claude CodeCodex
Handles every user message in assistant and autonomous modes: activates persona, runs mandatory 3-tier retrieval escalation (tree → codebase → web), detects and routes user directives (new aspirations, corrections, preferences), performs knowledge-freshness reconciliation, and captures interaction-level learning into…
Skill Claude CodeCodex
Resolves open hypotheses whose horizons have elapsed, extracts lessons from each outcome, calculates per-category accuracy stats, and generates a calibration report. Use whenever the aspirations loop hits the hypothesis-review cadence, the user says "how accurate are my predictions" or "review my hypotheses", or…
Skill Claude CodeCodex
Manages the framework SEED: build the seed manifest, PLANT (publish) the domain-stripped framework into a destination repo, verify, and diff. The seed is domain-free — planting it grows a NEW, empty environment (e.g. claude-mind) with no agents, no world, no learned state. To relocate a LIVING mind WITH its agents…
Skill Claude CodeCodex
Full sprint-planning exercise over the world + agent queues: runs /backlog-report, gathers the priority dashboard, probes every live agent's selector vantage against standing directives, sweeps hygiene lanes (duplicates, routing, priority mismatches, recurring health, reclaim-routed-work, zombie aspirations), applies…
Skill Claude CodeCodex
Creates or resumes an agent in reader (read-only), assistant (user-directed), or autonomous mode (perpetual loop), handling full initialization for new agents (Self, program, paths, aspirations, curriculum) and state transitions for existing ones. USER-ONLY — Claude must NEVER invoke /start. Fires only when the user…
Skill Claude CodeCodex
Stops the autonomous learning loop gracefully: writes stop-requested for Phase -1.4, waits for in-flight obligations (verify, state-update), consolidates session state via /aspirations-consolidate, drops the agent to assistant mode by default (or reader with --reader), and transitions state to IDLE. USER-ONLY — Claude…
Skill Claude CodeCodex
Surface portfolio-shape patterns the user should consider acting on. Tail consolidation pressure, work-class skew, aged aspirations, all-blocked aspirations. Hybrid skill — user-invocable AND agent-callable. Auto-fires every 50 completed goals via aspirations-precheck cadence.
Skill Claude CodeCodex
Relocates a LIVING mind — its agents, world, meta, and identity — onto another machine. Two modes. own-cloud: the repo (git) carries code + tracked agent identity and S3 carries world/meta, so the source side just prepares and prints a destination bring-up checklist (no archive). offline: packs a portable archive…
Skill Claude CodeCodex
Portable, read-only viewer for the knowledge tree using only built-in file tools — no framework scripts required. Use whenever the agent needs to browse tree structure without triggering retrieval counters, when operating in reader mode, during observer sessions that must stay side-effect-free, or when debugging tree…
Skill Claude CodeCodex
Knowledge-tree operations entry point dispatching to sub-commands: read (node content), find (search), add (new node), edit (existing node), set (field update), decompose (split large node), split-overcap (verbatim split of a node too big to Read — MUST use for crit3/oversized nodes, never hand-split), distill…
Skill Claude CodeCodex
Scans verify-learning/SKILL.md for stale Check:/Bash: assertions whose targets (paths, phase headers, grep patterns) no longer exist in the codebase. Four lanes: L1 path references, L2 Phase/Step references inside SKILL.md targets, L3 grep-target patterns, L3b Grep Phase X for pattern variant. Skips negative…
Skill Claude CodeCodex
Runs post-test verification: checks the agent's state against a comprehensive checklist of learning artifacts (tree encodings, resolved hypotheses, reflections, skill return-protocols, guardrails, pattern signatures) and reports missing or drifted items. Use whenever the user says /verify-learning, after major…
Skill Claude CodeCodex
The simplified per-Body execution loop a forked WORKER Body runs (Mind/Body convergence Phase 2, asp-306). select -> claim -> execute -> RE-ENTER for the next work unit; parks (resumable) when work or the reducer is gone, closes only when a park expires. Records the status it judged for each unit through the shared…
Instructions file
Instructions for zkysar1/Claude-Mind, covering claude.md, project purpose, architecture, framework vs state split (4-tier architecture) and core design principle: no terminal state.