The Agent Skills Directory

[SAFE]: No malicious patterns such as obfuscation, credential theft, or unauthorized network access were found. The skill describes an intended workflow for background job management.
[PROMPT_INJECTION]: The skill implements a security boundary by instructing the agent to avoid passing raw natural language input directly into the evolution process. Instead, it mandates a validation step using separate tools (ooo interview and ooo seed) to produce structured, validated YAML input. Ingestion points: User-provided requests and lineage identifiers entering via seed_content and lineage_id arguments in SKILL.md. Boundary markers: The instructions explicitly state that raw natural-language input must flow through the validated Seed path first, preventing direct command execution from arbitrary text. Capability inventory: The skill manages background jobs using the ouroboros_ralph tool and monitors them using the job suite (wait, status, result, cancel) as defined in SKILL.md. Sanitization: Relies on requirement capture and seed generation tools to sanitize and validate user intent before triggering the iterative evolution process.

ralph