The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted content from multiple external systems, creating an indirect prompt injection surface.
Ingestion points: The skill ingests data from GitHub (PRs, commits), Slack (public channels), Jira/Confluence (tickets, project docs), and Google Docs as detailed in DATA-SOURCES.md.
Boundary markers: The instructions do not specify the use of clear boundary markers or delimiters (such as XML tags) to isolate untrusted external content when processing it or including it in the final report.
Capability inventory: The skill has access to the gh CLI tool (subprocess) and file system write permissions for producing the retrospective reports and the sprint-retro-evidence.json file. It also supports parallel agent spawning for data gathering tasks.
Sanitization: Although PRIVACY.md outlines a '5-gate privacy checklist' for filtering sensitive keywords and PII, the skill lacks specific sanitization or filtering for instructional patterns or command injections within the retrieved data.
[COMMAND_EXECUTION]: The skill interacts with the local environment using the GitHub CLI tool to fetch repository data.
Evidence: DATA-SOURCES.md specifies the use of gh pr list and gh api to gather information on PRs and user activity.
[DATA_EXFILTRATION]: The skill facilitates the transfer of gathered information to an external cloud storage service.
Evidence: FORMAT.md states that the HTML version of the generated report is uploaded to Amazon S3 as part of its standard delivery workflow.

sprint-retro