The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process potentially untrusted data from external sources such as project trackers and requirement documents to guide the agent's decision-making loop.
Ingestion points: As specified in the 'Method' section of SKILL.md, the agent reads project governance, governing requirements, and live trackers.
Boundary markers: The skill lacks explicit instructions to treat these external inputs as untrusted or to use delimiters to separate these data sources from system instructions.
Capability inventory: The skill allows for command execution, branch management, and pull request creation as detailed in the 'Dispatch' and 'Delivery' sections.
Sanitization: No instructions are provided for sanitizing, escaping, or validating the content of the ingested files.
[PERSISTENCE]: The skill establishes persistent state by creating and maintaining artifacts in ~/.agents/artifacts/<workspace>/. This enables the 'campaign' to resume across different sessions and execution environments by reading previously stored goals and status ledgers.
[COMMAND_EXECUTION]: The instructions require the agent to execute a variety of project-specific commands for tasks including lane management, implementation verification, and deployment integration.

sus-campaign