The Agent Skills Directory

[PROMPT_INJECTION]: The skill establishes a multi-agent communication channel through handoff files and issue comments, creating an attack surface for indirect prompt injection.
Ingestion points: The agent reads and processes data from .claude/hooks/.state/handoff.json and the output of various bd commands (e.g., bd comment, bd ready).
Boundary markers: There are no defined delimiters or instructions to ignore embedded commands within the handoff messages or issue context.
Capability inventory: The skill is granted access to the Bash and Write tools, which could be used to execute commands or modify the codebase based on instructions received from other agent instances.
Sanitization: The skill does not specify any validation, escaping, or filtering for the content ingested from the shared multi-agent state.

swarm-coordination