gathering-migration
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests codebase data and user instructions which are passed to subagents, creating a potential injection surface.
- Ingestion points:
SKILL.mdreceives the migration request, andreferences/conductor-dispatch.mdreceives the scoutedterritory_mapderived from codebase analysis. - Boundary markers: Variables
{migration_spec}and{territory_map}are interpolated into prompts inreferences/conductor-dispatch.mdwithout the use of strict delimiters or instructions to the model to ignore embedded directives. - Capability inventory: The workflow utilizes
gffor codebase searching,gwfor git operations, and executespnpm installand test commands via the shell. - Sanitization: No sanitization or filtering logic for the ingested content is defined in the instructions.
- [COMMAND_EXECUTION]: The skill instructs the agent to run shell commands for project maintenance and verification.
- Evidence: Gate checks in
SKILL.md(Phases 3 and 4) executepnpm installandgw dev ci --affected --fail-fast --diagnoseto verify the state of the codebase. - [EXTERNAL_DOWNLOADS]: The skill performs dependency resolution which involves downloading external assets.
- Evidence: The
pnpm installcommand is executed as part of the validation process to ensure the migration has not broken the dependency tree.
Audit Metadata