The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skills/asc-mcp.md file provides instructions to install the asc-mcp tool from a non-whitelisted GitHub user (zelentsov-dev) using the mint package manager. This represents an unverifiable dependency from an unknown source.
[INDIRECT_PROMPT_INJECTION]: The skill incorporates several workflows (skills/production-triage.md, skills/testflight-triage.md) that ingest untrusted data from external sources.
Ingestion points: Reads crash logs in .ips and .crash formats, and fetches issue data and feedback from the Sentry and App Store Connect APIs.
Boundary markers: The instructions do not specify the use of delimiters or warnings to ignore instructions potentially embedded within the ingested data.
Capability inventory: The skill utilizes the xcsym tool for crash analysis and the triage-analyzer agent for corpus triage. It also performs network operations to fetch data from aggregator APIs.
Sanitization: There is no mention of sanitizing or validating the content of the external logs or feedback before processing.
[COMMAND_EXECUTION]: The skill guides the user to execute shell commands for package installation (brew, mint) and for configuring the environment, which involves handling sensitive API credentials.

axiom-shipping