specs-kit-sync

Pass

Audited by Gen Agent Trust Hub on Sep 8, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from multiple sources within the repository to automate documentation updates and task generation.
  • Ingestion points: The skill reads content from functional specification files (YYYY-MM-DD--*.md), task files (tasks/*.md), decision-log.md, user-request.md, and the codebase graph (graphify-out/graph.json).
  • Boundary markers: None identified in the skill instructions to delimit or isolate content ingested from these external files.
  • Capability inventory: The skill performs file reads, file writes (updating specs and creating new task files), and invokes the graphify tool/skill.
  • Sanitization: The instructions do not specify any sanitization, validation, or escaping of the ingested markdown or JSON content before it is interpolated into new documentation or tasks.
  • [COMMAND_EXECUTION]: The skill executes local system commands and invokes external tools as part of its synchronization workflow.
  • Evidence: Performs file backups using the cp command in Phase 7.1.
  • Evidence: Triggers execution of the graphify skill (e.g., /graphify <project-root> --update) to refresh codebase metadata in Phase 3.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 8, 2026, 07:11 AM
Security Audit — agent-trust-hub — specs-kit-sync