openspec-implementation-loop
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes several shell commands including
git,gh(GitHub CLI),make, andgo testto manage code changes, synchronize specifications, and validate builds. These commands are integral to the skill's developer-focused workflow. - [INDIRECT_PROMPT_INJECTION]: The skill functions as an orchestrator that ingests data from potentially untrusted external sources and applies it to a context with significant capabilities.
- Ingestion points: The skill reads
contextFiles(Step 3), parses output fromopenspecCLI commands, and monitors external PR comments and reviews (Step 11). - Boundary markers: There are no explicit delimiters or instructions to ignore embedded commands within the ingested content defined in the orchestrator instructions.
- Capability inventory: The skill can perform file modifications (via implementation subagents), create git commits, push code to remote repositories (
git push), and execute arbitrary build or test commands. - Sanitization: The instructions do not specify any sanitization, validation, or filtering of the content retrieved from external files or PR interactions before processing.
Audit Metadata