The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes a local Node.js script scripts/call.mjs to perform its primary functions, including network requests to the vendor's API and processing payments on the Base blockchain.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external registries such as Crossref and ORCID. This data, which includes publication abstracts and researcher biographies, could potentially contain malicious instructions intended to influence the agent's behavior.
Ingestion points: External metadata is retrieved via the crossref.doi-metadata, crossref.search, and orcid.researcher-profile service calls defined in SKILL.md.
Boundary markers: The instructions do not define specific delimiters or instructions for the agent to ignore embedded commands within the retrieved metadata.
Capability inventory: The agent has the capability to execute local scripts (scripts/call.mjs) and perform network operations via curl to https://fittings.sh.
Sanitization: There is no mention of sanitization or filtering of the external content before it is processed by the agent.
[CREDENTIALS_SAFE]: The skill requires a private key for transaction signing (FITTINGS_PRIVATE_KEY). While handling cryptographic keys is sensitive, the skill adheres to safety best practices by instructing the use of environment variables rather than hardcoding credentials and provides explicit warnings to the agent against leaking the key in user-facing output.

fittings-papers