The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data by scanning codebase comments (TODO, FIXME, HACK, XXX) and git logs to suggest improvement tasks. This introduces a potential surface for indirect prompt injection where instructions hidden in code comments could influence agent behavior.
Ingestion points: SKILL.md (Phase 2: Task Selection) scans local source code files and git history.
Boundary markers: The skill does not define specific delimiters or instructions to ignore potential prompts within the scanned content.
Capability inventory: The skill can create directories, write multiple markdown files to the local file system, and execute the openspec CLI tool.
Sanitization: No sanitization or filtering of the content found in codebase comments is mentioned before the agent processes and presents it.
[COMMAND_EXECUTION]: The skill executes local system commands to verify environment readiness and perform workflow actions.
Evidence: SKILL.md executes openspec --version, git log --oneline -10, openspec new change, openspec instructions proposal, mkdir -p, and openspec archive across various phases of the onboarding process.

openspec-onboard