The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the GitHub CLI (gh) to fetch information about pull requests and issues from the repository to generate the changelog.
[COMMAND_EXECUTION]: The instructions include a shell script example that uses curl to send the generated changelog to a Discord webhook. This is provided as an optional feature for users to configure.
[DATA_EXFILTRATION]: The skill involves network communication with Discord, a well-known third-party service, to deliver the finalized changelog. This is the intended behavior of the skill and uses official service domains.
[PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection as it processes untrusted data from GitHub PR descriptions and issues.
Ingestion points: External data is pulled from GitHub PRs and issues via gh cli (referenced in the 'PR Analysis' section).
Boundary markers: The instructions use <change_log> tags for the final output but do not define specific delimiters or "ignore instructions" warnings for the processed content from GitHub.
Capability inventory: The agent can execute shell commands (gh, curl) which are necessary for its primary function but could theoretically be targeted if the input content is malicious.
Sanitization: There is no explicit sanitization step described for the external content before it is analyzed and summarized.

changelog