coordinator-init-mode
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONPERSISTENCE
Full Analysis
- [COMMAND_EXECUTION]: The skill executes
git config user.nameto retrieve the developer's name for attribution within the team structure. It contains an explicit safety instruction to avoid readinguser.emailto protect user privacy. - [PERSISTENCE]: Modifies the repository's
.gitattributesfile to registerunionmerge drivers for squad-specific files. This ensures that log files and decision records persist correctly across branch merges without manual conflict resolution. - [INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided descriptions of the project and tech stack to populate the initial history and charters of generated agents.
- Ingestion points: User input provided during Phase 1 (SKILL.md).
- Boundary markers: The skill does not define specific delimiters for separating user-provided context from agent instructions in the generated
.squad/agents/*/history.mdfiles. - Capability inventory: The skill performs file system writes to the
.squad/directory and modifies the root.gitattributesfile. - Sanitization: There is no explicit sanitization step described for user-supplied strings before they are seeded into agent charters.
Audit Metadata