The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute a shell script located at ~/.claude/scripts/org-releases.sh to collect package data and redirect the output to a temporary file.
[DYNAMIC_EXECUTION]: As part of Phase 3 (Prepare releases), the skill runs R CMD check for R packages or equivalent tests for Julia packages. These testing tools execute the code within the repositories being processed, which constitutes dynamic execution of local code.
[INDIRECT_PROMPT_INJECTION]: The skill processes data ingested from /tmp/org-releases.json. This file contains data derived from external repositories, creating a surface for potential indirect prompt injection if the package metadata is maliciously crafted.
Ingestion points: The agent reads and classifies data from the /tmp/org-releases.json file generated in the helper script step.
Boundary markers: No specific boundary markers or instructions to treat the JSON content as untrusted data are present.
Capability inventory: The skill possesses the capability to execute shell scripts, perform git operations, create GitHub PRs via the gh CLI, and run package-level tests.
Sanitization: There is no evidence of sanitization or schema validation performed on the ingested JSON data before it is used to determine the release status and prepare PRs.

org-releases