The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses subprocess.run to execute git and grep commands for codebase analysis and staleness checking. While it does not use shell=True, it interacts with the local file system based on patterns extracted from documentation.
[EXTERNAL_DOWNLOADS]: The verify_claims.py script contains functionality to perform HTTP HEAD requests to external URLs found in documentation. This feature is optional and must be enabled via the --check-urls flag, but it allows the agent to communicate with arbitrary external domains.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it ingests and processes untrusted markdown files from the project.
Ingestion points: The skill scans all markdown files in the project directories to extract claims.
Boundary markers: There are no delimiters or specialized instructions to prevent the agent from obeying instructions embedded within the documentation it is verifying.
Capability inventory: The skill has access to the local file system, can execute specific shell commands, and can make network requests.
Sanitization: Input for the grep command is sanitized using re.escape(), but the content processed by AI subagents for behavioral and dependency verification lacks robust sanitization or validation.

doc-claim-validator