The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill requires the user to provide a sensitive cryptographic private key through the FITTINGS_PRIVATE_KEY environment variable. While the documentation advises using a throwaway wallet, the key is maintained in plaintext within the environment, which is a significant exposure risk in an AI agent context.
[EXTERNAL_DOWNLOADS]: The SKILL.md instructions include a command to download the client script (call.mjs) directly from the vendor's domain (https://fittings.sh/skill/call.mjs) using curl if the local file is missing.
[COMMAND_EXECUTION]: The skill provides explicit shell commands for the agent or user to execute, involving environment variable exporting and Node.js script invocation.
[INDIRECT_PROMPT_INJECTION]:
Ingestion points: The skill processes CVE identifiers provided as input and data returned from the fittings.sh API (SKILL.md, call.mjs).
Boundary markers: No explicit delimiters or instructions are used to prevent the agent from interpreting instructions that might be embedded in the external API responses.
Capability inventory: The skill possesses network access (fetch) and filesystem read access (call.mjs).
Sanitization: Input processing is limited to type coercion, with no sanitization of the content to prevent injection attacks.
[DYNAMIC_EXECUTION]: The call.mjs script performs a runtime integrity check by hashing its own source code and comparing it to a value fetched from the vendor's remote server. It also uses dynamic import() calls to load cryptographic and network libraries.

kev-status