The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides instructions to execute a PowerShell script located at .github/muscles/brain-qa.ps1 for automated repository audits. This includes 31 validation phases covering synapse targets, schema formats, and file synchronization between master and heir contexts.
[PROMPT_INJECTION]: The skill contains an indirect prompt injection surface as it is designed to perform semantic and logic validation on repository files matching patterns like **/*synapse*, **/*skill*, and **/*trigger*. Instructions embedded in these files could potentially attempt to influence the agent's semantic audit or logic validation results.
Ingestion points: Repository files identified by the applyTo patterns in SKILL.md.
Boundary markers: Absent. No specific delimiters are defined to separate untrusted file content from analysis instructions.
Capability inventory: The associated brain-qa.ps1 script performs file system operations, including reading file hashes and synchronization (writing/renaming files) in Phases 5, 7, 8, and 13.
Sanitization: No explicit sanitization or validation of the ingested file content is mentioned beyond the automated script checks.

Brain QA