The Agent Skills Directory

[COMMAND_EXECUTION]: The skill interacts with the diffowl CLI (e.g., diffowl findings, diffowl findings fix, diffowl findings dismiss) to update the status of code issues. It also executes repository-specific verification commands, such as pnpm run test, to validate fixes before they are recorded in the lifecycle database.
[DATA_EXPOSURE]: The skill reads analysis data and reports from the .diffowl/reviews/ directory. This is standard behavior for managing the tool's findings and does not involve accessing sensitive system or user configuration files.
[INDIRECT_PROMPT_INJECTION]: The skill ingests findings that could be influenced by code analysis of external contributions. The risk is mitigated by explicit instructions requiring the agent to verify all findings against the current codebase and project history.
Ingestion points: Findings are read from .diffowl/reviews/*.md and the output of diffowl findings.
Boundary markers: No explicit delimiters or boundary markers are used for the ingested content.
Capability inventory: The skill has the ability to execute diffowl CLI commands and arbitrary verification commands defined by the repository's workflow.
Sanitization: The skill relies on a verification step where the agent must confirm the finding's validity against the actual source code before taking action.

diffowl-resolve