The Agent Skills Directory

[SAFE]: No security threats were identified during the analysis. The skill implements strong boundary controls and hard stops to prevent unauthorized file edits and ensuring strict adherence to canonical approval protocols.
[COMMAND_EXECUTION]: The skill utilizes local CLI commands br (beads-cli) and bv (beads-viewer) to provide evidence of bead graph readiness. These are internal tools used for state verification and do not involve remote code execution or shell injection risks.
[DATA_EXPOSURE]: The skill reads and writes to local project state files such as STATE.json, readiness-record.json, and approval-record.json. This data is focused on task metadata and readiness rationale, with no evidence of accessing sensitive system files or credentials.
[SAFE]: The skill processes external artifacts like PLAN.md and CONTEXT.md to classify readiness. While this represents an indirect input surface, the instructions specifically command the agent to prioritize outputs from the br tool over prose descriptions, mitigating risks from potentially malicious content in those files.

beo-validate