The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to execute local utility scripts, specifically append-loop-result.ts and manifest-sync.ts, using the bun runtime. These scripts are located in a variable-defined directory (${SKILLS_ROOT}) and are used for workspace synchronization and results ledger maintenance. This is standard functionality for the author's workflow.
[PROMPT_INJECTION]: The skill processes untrusted external data, creating an indirect prompt injection surface.
Ingestion points: External evidence including analytics exports, experiment reports, and qualitative notes are ingested and processed by the Metric Ingest Agent and Diagnosis Agent in agents/metric-ingest-agent.md and agents/diagnosis-agent.md respectively.
Boundary markers: There are no explicit delimiters or instructions provided to the agents to ignore or escape instructions that might be embedded within the external data sources.
Capability inventory: The skill has significant permissions including Read, Write, Edit, and Bash. These capabilities are used to generate evaluation reports and update the loop ledger via shell commands.
Sanitization: The skill lacks explicit validation or sanitization mechanisms for the data retrieved from external sources before it is used to generate filenames or shell command arguments in the orchestrator.

lp-eval