The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes legitimate shell commands for development workflows, such as git operations and running project-specific tests.
Evidence: Use of git status and git diff to identify changes, and execution of unit/integration tests within Docker containers as defined in project rules.
[INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection as it ingests untrusted data from the repository and task definitions.
Ingestion points: The skill reads source code, migration files, harness rules, and task cards (task-*.md) provided in the repository.
Boundary markers: The skill employs structured markdown headers (e.g., ### Contract, Validation criteria) and explicit file loading protocols to manage data context.
Capability inventory: The skill has the ability to read and write to the file system, execute git commands, and run shell-based tests.
Sanitization: The risk is mitigated by a mandatory review pipeline involving ns-reviewer and ns-judge, requiring human approval or a high quality score (10/10) before tasks are finalized.
[SAFE]: No malicious patterns such as credential theft, data exfiltration, or persistence mechanisms were found. The skill follows established vendor patterns for nextstage-brasil components.

ns-coder