Peleke Sengstacke

Security-scanned agent skills from the brunnr registry. Every skill passes a 7-class deterministic security scanner before distribution.

Score agent-facing tool descriptions 0-5 on discoverability. Takes a tool description, explains failures against 5 criteria, and rewrites it to pass.

Run a structured AX Interview on agent-tool sessions. Gricean maxim analysis, implicature detection, CoT faithfulness check, Pragmatic Coherence Score with ranked fixes.

Turn a braindump into tomorrow's runlist. Input: unstructured text + optional previous runlist. Output: 4-quadrant Obsidian markdown (~800 tokens) with entry points on high-activation items, carry-forward tags, kill reasoning, and RUNLET_SUMMARY JSON for agent pings. ND-adapted: classifies by activation energy, not urgency.