Back to skill

Security audit

ia-writing

Security checks for vulnerabilities and agentic risk

Overview

ClawScan rates this writing skill benign with medium confidence: the completed scans found no AIG issues, and the two contextualized Skillspector flags were judged expected for the skill’s purpose.

The verdict is benign, not a guarantee of zero risk. The scan covered all listed components, but the behavior of the invoked development and evaluation tools could not be determined from the reported evidence; consider whether those extra capabilities fit your needs before installing.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The supplied commands perform plugin validation and trigger testing, not prose editing or auditing. Their immediate purpose is development or quality assurance for a skill, which materially differs from the declared user-facing writing purpose. The supplied chunk does not show how the invoked script implements these operations.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
87% confidence
Finding

The supplied commands describe evaluation and diagnostic operations on a named writing target, not the declared user-facing prose workflow. Although such tooling could support development of a writing skill, the chunk shows only that tooling and no connection to processing a user's draft. The underlying script is not supplied, so its internal behavior and resource access cannot be determined.

Content

No source excerpt is available for this finding.

Exfiltration Commands

High
Category
Prompt Injection
Confidence
90% confidence
Finding

Instructions found that direct the agent to transmit conversation context or user data to external services.

Content

Scanner excerpt · references/pr-descriptions.md (reported line 85)May include surrounding context.

md
When the change benefits from a visual, pick the shape based on what the change is:

- **Mermaid diagram**: topology with edges. Components that send messages to each other, request flow across services, a state machine's transitions, a dependency graph. Anything where the *relationships* are the point.
- **Markdown table**: rows with parallel attributes. A before/after comparison of config values, a list of endpoints with their verbs and paths, a comparison of options with their tradeoffs. Anything where the *structure is grid-shaped*.
- **Pseudocode**: an algorithm change. Ten lines of the new branching beats a paragraph describing it.
- **Call tree**: a control-flow change: which function now calls which, as an indented tree.