The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: Fetches a service manifest (agent-manifest.json) from the vendor's GitHub Pages domain (cristianmoroaica.github.io) to locate the active API endpoint.
[COMMAND_EXECUTION]:
Dynamic Service Resolution: The skill determines its primary API endpoint at runtime by reading an external JSON file, allowing the service location to change without updating the skill code.
Financial Transactions: Instructs the agent to participate in a blockchain-based payment protocol (x402 challenge/Bazaar method) involving specific transaction parameters, such as a USDC contract address on the Base mainnet.
[DATA_EXFILTRATION]: Transmits metadata of the projects being audited (URL-encoded repository URLs and skill paths) to a dynamically resolved external API.
[PROMPT_INJECTION]:
Ingestion points: Processes audit data and follows evidence_url links provided by the external 'SkillVerdict' API (SKILL.md).
Boundary markers: Lacks explicit delimiters or instructions to ignore embedded commands within the content retrieved from the audit service or evidence links.
Capability inventory: Performs network requests to interact with external APIs and resolves secondary links based on untrusted service output.
Sanitization: Does not specify validation or sanitization for the content of the evidence_url or the audit report findings before processing.

preflight-agent-skills