The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to analyze and verify untrusted artifacts, creating a potential surface for indirect prompt injection attacks.
Ingestion points: The skill processes external artifacts, implementations, and source files identified during the investigation phase.
Capability inventory: The agent is instructed to use search_files, read_file, ls, find, hugo, and curl during the verification process.
Sanitization: The instructions do not specify any validation or sanitization for the content of the artifacts being processed.
Boundary markers: While the skill uses a structured protocol, it does not mandate specific boundary markers to isolate untrusted data from the agent's instruction context.
[COMMAND_EXECUTION]: The methodology requires the execution of shell commands to validate criteria.
The skill instructs the agent to use curl to verify that links within an artifact resolve correctly, which involves initiating network requests to arbitrary external URLs.
It uses the hugo command to verify that static site builds exit without error.
Common file system utilities such as ls and find are utilized to locate duplicate files or verify directory structures.

verification-methodology