learn-with-docs

Pass

Audited by Gen Agent Trust Hub on Oct 12, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill uses npx to fetch the @sourcemeta/jsonschema package from the NPM registry during the documentation verification process. This involves downloading external code at runtime to perform schema validation (referenced in references/verify-claims.md).
  • [COMMAND_EXECUTION]: The skill instructions direct the agent to execute several shell commands for repository analysis and documentation maintenance. This includes git log, git blame, git show, and gh issue list for gathering context from the repository history and associated issues (referenced in SKILL.md and references/verify-claims.md).
  • [DYNAMIC_EXECUTION]: The use of npx -y @sourcemeta/jsonschema validate executes third-party code downloaded at runtime to verify the structure of documentation frontmatter (referenced in references/verify-claims.md).
  • [INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted data from external sources, presenting an attack surface for indirect prompt injection.
  • Ingestion points: The skill reads existing documentation in docs/solutions/, retrieves data from git commit messages and logs, and fetches issue/PR content via the GitHub CLI (referenced in SKILL.md and references/verify-claims.md).
  • Boundary markers: The instructions do not mandate the use of strict delimiters or isolation markers when the agent reads or interpolates these external data sources into its drafting process (referenced in SKILL.md).
  • Capability inventory: The agent has the capability to write to the filesystem (docs/solutions/, CONCEPTS.md, AGENTS.md) and execute shell commands (git, gh, npx, awk, find), which could be leveraged if malicious instructions were embedded in the ingested data (referenced in SKILL.md and references/verify-claims.md).
  • Sanitization: There are no specific instructions provided to escape or sanitize the content retrieved from version control or external issues before it is used in documentation generation or as arguments for commands (referenced in SKILL.md).
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 12, 2026, 02:59 PM