GLM-OCR
Analysis
This OCR skill appears purpose-aligned, but users should notice that selected files or URLs are sent to GLM/Zhipu and an API key is stored locally.
Findings (3)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Checks for instructions or behavior that redirect the agent, misuse tools, execute unexpected code, cascade across systems, exploit user trust, or continue outside the intended task.
**ONLY use GLM-OCR API** ... **NEVER offer alternatives** ... **IF API fails** - Display the error message and STOP immediately
The skill gives strict tool-choice and stopping instructions. They are aligned with keeping OCR work on the GLM API path, but they could prevent the agent from offering alternate approaches when a user might want them.
Checks whether tool use, credentials, dependencies, identity, account access, or inter-agent boundaries are broader than the stated purpose.
env_vars = {"ZHIPU_API_KEY": api_key} ... f.write(f"{key}={value}\n")The setup helper writes the API key into a local .env file. This is expected for an API integration, but it creates a persistent local credential.
Checks for exposed credentials, poisoned memory or context, unclear communication boundaries, or sensitive data that could leave the user's control.
DEFAULT_API_URL = "https://open.bigmodel.cn/api/paas/v4/layout_parsing" ... data = path.read_bytes() ... resp = requests.post(api_url, json=payload, headers=headers, timeout=timeout)
Local files are read, encoded into the API payload, and sent to the external GLM OCR endpoint. This matches the OCR purpose, but private documents could be transmitted to the provider.