The Agent Skills Directory

[PROMPT_INJECTION]: The skill facilitates a workflow where the agent ingests external data from pull requests, establishing a surface for indirect prompt injection. Ingestion points: The agent is instructed to read pull request metadata (titles, bodies, and branch names) via the multica issue pull-requests command and custom properties via multica issue property list (SKILL.md). Boundary markers: The instructions do not define specific delimiters or instructions for the agent to disregard potentially malicious embedded commands within the ingested text. Capability inventory: The agent has access to state-mutating and information-gathering tools including multica, git, and gh via the bash tool (SKILL.md). Sanitization: The documentation lacks instructions for validating, escaping, or sanitizing the content retrieved from external PR sources before it is processed by the agent.

multica-working-on-issues