Orchestrator PM

• Risk level: Low
• Analyzed: Mar 31, 2026, 5:25 AM
• Findings: 0
• Version: 2.0.0

Overview

The skill's instructions, required files, and behavior are internally consistent with a delegating/project-orchestration role; it requests no credentials or installs and only reads/writes workspace artifacts as described.

This skill appears internally coherent for orchestrating and delegating work, but review before enabling: 1) Inspect workspace files (CONTEXT.md, TASKS.md, .learnings/, etc.) for secrets or sensitive data — the skill reads/writes these files and will persist insight tweets and risk register entries. 2) Confirm you trust the downstream skills it will invoke (spawner, recorder, verifier, publisher); those skills will receive workspace context. 3) If you want a human check before delegations, ensure the 'human-escalation' path remains enabled or modify the instructions so gating decisions require explicit human approval. 4) Test in a sandbox workspace first to verify produced JOB_BOARD.md and NEXT_AGENT.md look as expected and do not leak private information.