core-planner

Fail

Audited by Gen Agent Trust Hub on Feb 20, 2026

Risk Level: HIGHCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION] (HIGH): The skill's YAML frontmatter defines 'pre' and 'post' hooks that execute shell commands using the $TASK variable. Evidence: 'echo "🎯 Planning agent activated for: $TASK"' and 'memory_store "planner_start_$(date +%s)" "Started planning: $TASK"'. This allows an attacker to execute arbitrary shell commands if the task name contains shell metacharacters such as backticks or semicolons.
  • [PROMPT_INJECTION] (LOW): The skill is susceptible to indirect prompt injection through task descriptions. 1. Ingestion points: Untrusted task descriptions entering via the $TASK variable and planner input. 2. Boundary markers: No delimiters or ignore-instruction warnings are present. 3. Capability inventory: The skill has file read (Read, Glob, Grep) and file write (TodoWrite) capabilities. 4. Sanitization: No sanitization or validation of the input task content is performed.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Feb 20, 2026, 03:44 PM