The Agent Skills Directory

[COMMAND_EXECUTION]: The skill facilitates the generation and execution of curl commands to interact with the TypeSafe Jev HTTP API. The implementation follows security best practices by disabling shell tracing (set +x) and passing sensitive authorization headers via standard input (--header @-) to ensure credentials do not appear in process listings or shell history.
[INDIRECT_PROMPT_INJECTION]: The skill acts as an interface for processing external data through a remote AI model, creating a potential surface for indirect prompt injection.
Ingestion points: External data is ingested into the state field of the API request as defined in SKILL.md and references/http.md.
Boundary markers: The skill uses a structured JSON schema to separate the data (state), the model's tasks (instructions), and the evaluation rules (criteria).
Capability inventory: The primary capability is executing network requests via the curl binary as described in references/http.md.
Sanitization: The skill mitigates common injection risks by recommending the use of JSON files or quoted heredocs for data transport, which prevents shell metacharacter interference.

jev-curl