Back to skill

Security audit

a2awire-spend-guard

Security checks for vulnerabilities and agentic risk

Overview

ClawScan rates this skill suspicious because its bug-reporting instructions may expose sensitive purchase details or credentials; the scan is partial, so the evidence is incomplete.

Review the skill’s data-sharing and bug-reporting instructions before installing. Avoid sending reports with real purchase details, identifiers, or credentials; the scan was partial because some referenced paths were missing, so it cannot provide complete assurance.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (6)

External Script Fetching

High
Category
Supply Chain
Confidence
90% confidence
Finding

Remote code is downloaded and executed. This bypasses code review and could introduce malicious code.

Content

Scanner excerpt · SKILL.md (reported line 222)May include surrounding context.

md
when you already hold one — your guard must bind to **your** A2AWire
  identity's key (the `a2awire-join` skill), or the rules and ledger it
  keeps are some other identity's.
- Do NOT pipe any script to a shell (`curl ... | sh`). Download, read,
  verify, then run.

## Report problems

External Script Fetching

High
Category
Supply Chain
Confidence
90% confidence
Finding

Remote code is downloaded and executed. This bypasses code review and could introduce malicious code.

Content

Scanner excerpt · SKILL.md (reported line 222)May include surrounding context.

md
when you already hold one — your guard must bind to **your** A2AWire
  identity's key (the `a2awire-join` skill), or the rules and ledger it
  keeps are some other identity's.
- Do NOT pipe any script to a shell (`curl … | sh`). Download, read,
  verify, then run.

## Report problems

Skill Enumeration

Medium
Category
Agent Snooping
Confidence
80% confidence
Finding

Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.

Content

Scanner excerpt · BUILD-NOTES.md (reported line 19)May include surrounding context.

md
/api/v1/content/muse-spend-guard.md`, 2026-09-20) and cold-start
companion (`GET /api/v1/content/spend-guard-cold-start.md`,
2026-09-21); the commissioned local skill
(`~/workspace/custom-skills/local-spend-guard/SKILL.md` + its helper
source, which is Koda's production client for this lane). Koda-specific
test-identity details and all rule values from the local setup were
deliberately NOT carried into the skill.

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% 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 60)May include surrounding context.

bash
# initialize
curl -sS https://a2awire.com/mcp/connectors/spend-guard/http \
  -H "Authorization: Bearer <api_key>" \
  -H "Content-Type: application/json" \
  -H "Accept: application/json, text/event-stream" \

Bug-reporting instructions require exact request/response data without warning that it may disclose sensitive purchase information.

Medium
Category
Output Handling
Confidence
94% confidence
Finding

Asking for exact request and response data can cause an agent to include merchant names, purchase descriptions, transaction identifiers, and private spending rules in feedback. Although the service already processes these data, submitting them to a feedback channel introduces additional retention and access exposure without requiring minimization or user consent.

Content

No source excerpt is available for this finding.

Bug-reporting instructions request exact authenticated requests without requiring credential redaction.

Medium
Category
Data Exfiltration
Confidence
83% confidence
Finding

Exact authenticated requests may include Authorization or X-API-Key headers, exposing the identity's credential through a feedback submission. The earlier prohibition on putting keys in messages mitigates this risk, but the reporting instruction creates conflicting guidance without explicitly requiring sanitized requests; a leaked key could permit unauthorized ledger access or rule changes.

Content

No source excerpt is available for this finding.