10 actions in 3 pieces of software · ranked by words, meaning and measured trust · 711 ms
- 1graphnosis-appfound in code, not yet checkedno passport yet
MCP servernehloo-interactive/graphnosis-app
- set_skill_role
DETERMINISM — Deterministic: patches the skill's training-metadata comment; no LLM, no retrain.
- promote_import
DETERMINISM — Deterministic: moves reviewed sources out of quarantine via the trusted in-app moveSource and lifts the quarantine flag. No LLM.
- remember
DETERMINISM — Deterministic: saving the same note produces the same memory; no LLM and no randomness are involved, and every write is auditable. Add a note to the user's personal Graphnosis memory so it persists across sessions. ROUTING: When the note belongs to a topic / project / collection (not a
- set_skill_autonomy
DETERMINISM — Deterministic: persists ONE skill's execution-autonomy override in engram metadata; no LLM.
- set_engram_autonomy
DETERMINISM — Deterministic: persists the engram's execution-autonomy dial in metadata; no LLM.
- set_skill_roles
DETERMINISM — Deterministic: batch-patches training-metadata Role lines; no LLM.
- create_engram
DETERMINISM — Deterministic: creates an empty engram with a template + display name. No LLM, no recall.
- edit
DETERMINISM — Conditional. Deterministic by default: with no Local LLM and no Neural Network enabled, `edit` deterministically supersedes the single closest-matching memory with the new content (or records it as a new memory when nothing matches) — reproducible, no randomness, fully auditable. The o
- set_skill_role
- add-command
add-command — an installable skill for AI agents.
- add-command
- 3ao-lensfound in code, not yet checkedno passport yet
MCP servercredentum/ao-lens
- check_determinism
Check for replay-safety violations (os.time, math.random, io operations)
- check_determinism
Labels are the instruments' verdict and are never changed by ranking. Agents and robots get this same answer as JSON at /api/v1/find?q=…&answer=1 and through the MCP server.