The Agent Skills Directory

[COMMAND_EXECUTION]: The skill enables the execution of shell commands such as pnpm for testing and scripts/committer for version control operations.
[REMOTE_CODE_EXECUTION]: It utilizes the blacksmith CLI to run test suites on remote CI/CD infrastructure.
[EXTERNAL_DOWNLOADS]: The skill uses pnpm install to manage project dependencies, which involves fetching packages from the npm registry.
[PROMPT_INJECTION]: The agent is instructed to read various local AGENTS.md files as part of its workflow, representing an indirect prompt injection surface where documentation content could influence agent logic.
Ingestion points: Multiple AGENTS.md files including src/agents/AGENTS.md, src/channels/AGENTS.md, src/plugins/AGENTS.md, src/gateway/AGENTS.md, test/helpers/AGENTS.md, test/helpers/channels/AGENTS.md, and src/infra/outbound/AGENTS.md.
Boundary markers: The instructions do not specify the use of delimiters or sanitization when reading these files.
Capability inventory: The skill possesses extensive capabilities including file system modification, command execution, and remote job submission.
Sanitization: No content validation or escaping is applied to the data read from the local documentation files.

openclaw-test-performance