Trust Assessment
parallel-task received a trust score of 82/100, placing it in the Mostly Trusted category. This skill has passed most security checks with only minor considerations noted.
SkillShield's automated analysis identified 2 findings: 0 critical, 1 high, 1 medium, and 0 low severity. Key findings include Indirect Prompt Injection via Untrusted Plan Files, Uncontrolled Parallel Subagent Execution.
The analysis covered 4 layers: Manifest Analysis, Static Code Analysis, Dependency Graph, LLM Behavioral Safety. All layers scored 70 or above, reflecting consistent security practices.
Last analyzed on October 1, 2026 (commit 439cd3a9). SkillShield performs automated 4-layer security analysis on AI skills and MCP servers.
Layer Breakdown
Behavioral Risk Signals
Security Findings2
| Severity | Finding | Layer | Location | |
|---|---|---|---|---|
| HIGH | Indirect Prompt Injection via Untrusted Plan Files The skill parses external plan files (which may be untrusted or modified by an attacker) and directly interpolates their contents (descriptions, acceptance criteria, and validation steps) into the prompts of newly spawned subagents. If a plan file contains malicious instructions or shell commands in the 'Validation' or 'Description' fields, the subagents will execute them, potentially leading to arbitrary command execution, data exfiltration, or file destruction. Sanitize and validate all inputs extracted from plan files before passing them to subagents. Avoid executing arbitrary validation commands specified in the plan file without explicit user confirmation. Implement strict boundaries on what subagents can execute, and do not automatically run validation scripts/commands from untrusted plans. | LLM | SKILL.md:37 | |
| MEDIUM | Uncontrolled Parallel Subagent Execution The skill automatically spawns and executes multiple parallel subagents in a loop until the entire plan is completed, without requiring user confirmation between tasks or phases. This can lead to rapid, uncontrolled execution of destructive actions or excessive resource/API credit consumption if a plan is malicious or contains a loop. Require explicit user confirmation before launching each batch of parallel subagents or before executing tasks that modify the workspace. | LLM | SKILL.md:91 |
Scan History
Embed Code
[](https://skillshield.io/report/c7cc573105d0e9d3)
Powered by SkillShield