kata-resume-work
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill workflow defined in
references/resume-project.mdutilizes standard shell commands such asls,cat, andfindto inspect project artifacts and status. These operations are limited to the local filesystem and specifically target files within the.planning/metadata directory. - [INDIRECT_PROMPT_INJECTION]: The skill processes data from project-local files (
STATE.md,PROJECT.md,ROADMAP.md, and.continue-here*.md) to reconstruct session context. This represents a potential surface for indirect prompt injection if these files are sourced from an untrusted repository, although the skill uses this data purely for informational status reporting and workflow navigation. - Ingestion points: Reads project-local files
STATE.md,PROJECT.md,ROADMAP.md, and checkpoint files (.continue-here*.md) within thereferences/resume-project.mdworkflow. - Boundary markers: Absent; the skill relies on the agent's ability to interpret markdown structure to extract specific fields.
- Capability inventory: Performs file read/write operations within the
.planning/directory and executes local shell commands for status checks. - Sanitization: None; the skill extracts specific values (e.g., 'Project Reference', 'Phase') and presents them to the user.
Audit Metadata