The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute several local shell commands including make test, git diff, and uv run pytest. These commands are standard for software development and necessary for identifying and verifying test fixes within a local environment.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted data from external files and tool outputs.
Ingestion points: The agent reads output from make test and git diff, as well as the contents of source code and test files to identify root causes of failures.
Boundary markers: The instructions do not specify any delimiters or special handling to differentiate between the data being processed (e.g., test error messages or code comments) and instructions to the agent.
Capability inventory: The agent is granted the ability to modify local files using the Edit tool and execute further shell commands to verify fixes.
Sanitization: There is no evidence of sanitization or filtering of the ingested content, which could allow maliciously crafted test failures or code comments to influence the agent's behavior during the fixing process.

test-fixing