dstack-beads-plan-feature

Pass

Audited by Gen Agent Trust Hub on Aug 29, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes local shell commands using the dstack and bd (Beads) CLI tools. It passes variables such as <title>, <slug>, <selector>, and <id> directly into shell strings. While the instructions explicitly advise using --body-file to handle multiline data safely and avoid shell syntax issues, single-line variables remain potential vectors for command injection if the underlying agent platform does not perform automatic sanitization.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data to perform its task.
    • Ingestion points: The skill is instructed to inspect architecture diagrams, product documentation, source code, and test files in the SKILL.md workflow.
    • Boundary markers: No explicit delimiters or boundary markers are defined for the analysis of these external files.
    • Capability inventory: The skill has the capability to execute shell commands via the dstack and bd binaries and create/write temporary Markdown files.
    • Sanitization: The skill includes a mitigation strategy for multiline content by using the --body-file flag to persist intent as data rather than shell syntax, though it does not specify sanitization for metadata fields like titles or selectors.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 29, 2026, 12:02 AM
Security Audit — agent-trust-hub — dstack-beads-plan-feature