ClawHub
Back to skill

Security audit

SpecStream: Read a Request

Security checks for vulnerabilities and agentic risk

Overview

This skill sends requests to the configured SpecStream service for interpretation without directing the agent to start data-gathering jobs.

Compilation sends your request to SpecStream even though it does not start subsequent data gathering. Use the connector's sign-in, omit unnecessary confidential information, and review the service's permissions, retention, and sharing settings. The remote service implementation is outside this artifact review.

A.I.G

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

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

External Transmission

Medium
Category
Data Exfiltration
Confidence
50% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 9)May include surrounding context.

md
# Read the request

SpecStream is the live web data engine for teams and agents. Use the configured MCP server at `https://api.specstream.ai/mcp`.

Use `compile` to read the person's request. Say back what it found: the kind of thing, the conditions, the details to bring back and the quantity. Separate the conditions that must hold from the details that are nice to have. Use `edit_spec` for a deliberate change. Never drop a hard condition quietly, and never treat missing evidence as a no. Reading a request does not start any work.

The skill sends the person's request to SpecStream without a user-facing privacy disclosure.

Medium
Category
Output Handling
Confidence
92% confidence
Finding

The skill directs the agent to submit the person's request to a third-party compiler without requiring a user-facing explanation that compilation transmits the request externally. Requests may contain confidential information, and assurances that reading the request does not start work could obscure this initial disclosure. The configured connector and sign-in reduce concern about an unexpected destination, but do not establish informed consent for sensitive request content.

Content

No source excerpt is available for this finding.