The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructions direct the agent or user to download an external JavaScript file (call.mjs) from a remote server controlled by the vendor.
Evidence: curl -sO https://fittings.sh/skill/call.mjs in SKILL.md.
[REMOTE_CODE_EXECUTION]: The skill executes the previously downloaded external script using the Node.js runtime. This introduces a risk where the executed code is not audited as part of the skill itself and could be modified on the remote server.
Evidence: node call.mjs gov.congress-bill ... in SKILL.md.
[CREDENTIALS_UNSAFE]: The skill requires a blockchain private key for payment processing and instructs the user to export it as an environment variable (FITTINGS_PRIVATE_KEY). While it includes a warning to use a throwaway wallet, handling raw private keys in plaintext environments is a high-risk security practice.
Evidence: export FITTINGS_PRIVATE_KEY=0x... in SKILL.md.
[COMMAND_EXECUTION]: The skill relies on shell commands to perform its core functions, including downloading code and executing the runtime client.
Evidence: Use of curl, node, and export commands throughout the usage examples in SKILL.md.
[INDIRECT_PROMPT_INJECTION]: The skill ingests legislative data from a remote API which is then processed by the agent, creating a surface for potential injection if the source data is manipulated.
Ingestion points: https://fittings.sh/v1/gov/congress-bill (via call.mjs)
Boundary markers: None present in the skill instructions to separate external data from agent instructions.
Capability inventory: Shell command execution, file system access (writing call.mjs), environment variable access.
Sanitization: No sanitization or validation of the remote API response is specified.

fittings-gov-congress-bill