woostack-build

Pass

Audited by Gen Agent Trust Hub on Sep 7, 2026

Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes shell scripts and the GitHub CLI (gh) to manage project states, paginate issues, and verify contracts. These operations are performed within a scoped environment to automate repository-related tasks.
  • [DYNAMIC_EXECUTION]: Structural validation scripts (tests/test-provider-build-contract.sh, tests/test-provider-plan-contract.sh) use Bash heredocs to execute embedded Python code for internal consistency checks. This code is static and serves to enforce the skill's logic constraints during development and testing.
  • [EXTERNAL_DOWNLOADS]: The skill interacts with well-known service APIs and Model Context Protocol (MCP) servers for GitHub, Linear, and Plane to synchronize project artifacts. These connections are standard for the skill's stated purpose of project management mirroring.
  • [INDIRECT_PROMPT_INJECTION]:
  • Ingestion points: The skill reads external data, such as issue descriptions and project metadata, from GitHub, Linear, and Plane during synchronization phases.
  • Boundary markers: Explicit markers (e.g., <!-- woostack-spec-start -->) are used to delimit managed content and prevent confusion with human-authored text.
  • Capability inventory: The skill can perform mutations on external project boards and write local artifacts, but it requires explicit user approval steps (Stop here, Execute, Abandon) before final execution.
  • Sanitization: Instructions mandate treating unrelated human-authored content as untrusted evidence and emphasize manual verification for merging.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 7, 2026, 03:41 PM
Security Audit — agent-trust-hub — woostack-build