The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests data from external sources, including GitHub Pull Request bodies and issue descriptions from Jira and Linear. These inputs are not trusted and could potentially contain malicious prompts intended to influence the agent's behavior during the summary or publishing phases.
Ingestion points: Pull Request bodies (GitHub), Issue descriptions and summaries (Jira, Linear).
Boundary markers: The skill does not implement delimiters or specific instructions to the agent to disregard instructions found within the external data.
Capability inventory: The skill has capabilities to execute shell commands (via gh CLI), write to the local file system (saving to the vault), and perform network operations via MCP tools to publish to Notion, Confluence, and HackMD.
Sanitization: There is no evidence of data sanitization or filtering applied to the retrieved external content before it is interpolated into the release note templates.
[COMMAND_EXECUTION]: The skill utilizes the gh CLI and bash commands (mkdir, date) to perform its functions. It interpolates repository identifiers and version tags into these commands. While standard for this type of utility, it relies on the execution environment to prevent command injection from user-provided or tracker-provided metadata.

generate-release-notes