Back to skill

Security audit

AI Power Grid

Security checks for vulnerabilities and agentic risk

Overview

ClawScan labels this skill benign because the findings are considered consistent with its described AI Power Grid API use, though the scan was partial.

The verdict is benign with medium confidence, but benign does not mean risk-free. The skill describes sending requests to an external API, including requests that use an API key; review whether you trust that service and are comfortable using it. The scan was partial because a local reference could not be matched to a bundled artifact, so the evidence is incomplete.

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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (8)

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 17)May include surrounding context.

md
# AI Power Grid (AIPG)

A network of independent GPU operators behind one API at `https://api.aipowergrid.io/v1`. It is
OpenAI-compatible and billed per use from the account's credit. This skill is community-written and not
an official AIPG document.

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 77)May include surrounding context.

Find out what is online and what it costs (both public, no key needed):

bash
curl -s https://api.aipowergrid.io/v1/status/models   # name, type (text/image/video/audio), count = workers online
curl -s https://api.aipowergrid.io/v1/pricing         # price_book.models[].rates, e.g. per_image_usd

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 78)May include surrounding context.

bash
curl -s https://api.aipowergrid.io/v1/status/models   # name, type (text/image/video/audio), count = workers online
curl -s https://api.aipowergrid.io/v1/pricing         # price_book.models[].rates, e.g. per_image_usd

Use only models whose count is above 0. Send a non-default User-Agent: the API's edge refuses some default

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
80% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · SKILL.md (reported line 37)May include surrounding context.

md
Check for the environment variable `AIPG_API_KEY` (a key starts with `grid_`), or a saved profile:
`openclaw models auth list` shows `aipg:` entries.

If there is none, **do not try to make one and do not ask the user to paste a key into the chat.** Tell the
user:

1. Open https://console.aipowergrid.io/?callbackUrl=%2Fdashboard%2Fapi-key and sign in (Google, GitHub or a

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 77)May include surrounding context.

Find out what is online and what it costs (both public, no key needed):

bash
curl -s https://api.aipowergrid.io/v1/status/models   # name, type (text/image/video/audio), count = workers online
curl -s https://api.aipowergrid.io/v1/pricing         # price_book.models[].rates, e.g. per_image_usd

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 87)May include surrounding context.

Generate an image and save it:

bash
curl -s https://api.aipowergrid.io/v1/images/generations \
  -H "Authorization: Bearer $AIPG_API_KEY" -H "Content-Type: application/json" \
  -d '{"model":"z-image-turbo","prompt":"a small red lighthouse at dusk, flat illustration","n":1,"size":"1024x1024"}'

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 98)May include surrounding context.

Text, OpenAI chat format:

bash
curl -s https://api.aipowergrid.io/v1/chat/completions \
  -H "Authorization: Bearer $AIPG_API_KEY" -H "Content-Type: application/json" \
  -d '{"model":"qwen38-flash-next-nvfp4","messages":[{"role":"user","content":"Hello"}],"max_tokens":1024}'

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 109)May include surrounding context.

Credit left:

bash
curl -s https://api.aipowergrid.io/v1/account/credits -H "Authorization: Bearer $AIPG_API_KEY"

total_spendable_usd is what can be spent now; free is today's allowance.