The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill utilizes the gh CLI to clone external repositories from GitHub for read-only analysis, downloading content from untrusted external sources.\n- [COMMAND_EXECUTION]: Shell commands including mkdir, cd, gh repo clone, and grep are executed to manage the temporary workspace and search for process patterns within the cloned code.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted files from external repositories, such as journal logs and process definitions, which could contain adversarial instructions intended to influence the agent's analysis.\n
Ingestion points: External repository files including run.json, inputs.json, and the journal event stream (journal/*.json) are loaded into the agent's context for review.\n
Boundary markers: The instructions do not define delimiters or specific warnings to ignore instructions that might be embedded within the analyzed data.\n
Capability inventory: The skill is capable of writing retrospective reports to the local file system and executing shell commands for repository inspection.\n
Sanitization: There is no mention of sanitizing or filtering the content of the external files before the agent processes them for retrospective generation.

retrospect-external-babysitter-run