antigravity-image-gen
Verdict safe, but 2 findings (2 medium), mainly data exfiltration, unauthorized tool use. Most recent scan 2026-02-05.
MODERATE RISK - This skill presents several security concerns primarily around credential exposure and lack of transparency. The hardcoded OAuth credential path (~/.clawdbot/agents/main/agent/auth-profiles.json) creates a predictable target for credential theft. The missing 'allowed-tools' declaration enables unrestricted tool access, which could be exploited if the generate.js script (not provided for analysis) is malicious. The description claims to use an 'internal Google API' without verification, which could be social engineering. However, without access to the actual generate.js implementation, we cannot confirm data exfiltration, command injection, or other critical threats. The skill requires OAuth credentials and network access but lacks transparency through missing license/compatibility metadata. RECOMMENDATION: Do not use this skill until: (1) generate.js source code is reviewed for malicious network calls, (2) allowed-tools restrictions are added, (3) credential handling is secured, and (4) API endpoint authenticity is verified.
Scanner's overall assessment (LLM analyzer)
Second scanner: NVIDIA SkillSpector
Independent static run, no language model, over the same files as the Cisco scan above. How the two scanners compare across the whole corpus: Two scanners, one corpus.
What this skill does
Generate images using the internal Google Antigravity API (Gemini 3 Pro Image). High quality, native generation without browser automation.
- Package
- 2 files, 7.3 KB
- SKILL.md
- 182 words, 7 sections
- Code
- JavaScript (1)
- Requires
- bins:
node
SKILL.md covers: Antigravity Image Generation · Prerequisites · Usage · Direct Script Execution · Arguments · Output · Troubleshooting
Findings
MEDIUM Hardcoded credential path exposes authentication data Data Exfiltrationllm
The skill instructions explicitly reference a hardcoded credential file path '~/.clawdbot/agents/main/agent/auth-profiles.json' containing Google Antigravity OAuth credentials. This creates a predictable target for credential theft. If the skill's actual implementation (generate.js - not provided for analysis) reads this file, it could exfiltrate OAuth tokens to external endpoints. The instructions also reveal the internal authentication architecture, which aids attackers in targeting credential storage.
SKILL.md:9
**Google Antigravity OAuth Profile**: Must be present in `~/.clawdbot/agents/main/agent/auth-profiles.json`.MEDIUM Missing allowed-tools declaration enables unrestricted tool access Unauthorized Tool Usellm
The skill manifest does not declare 'allowed-tools', which means the agent has no restrictions on tool usage. Given that the skill requires Node.js execution and file system access (reading OAuth credentials, writing image files), the absence of tool restrictions creates risk for tool exploitation. An attacker could modify the generate.js script (not provided) to use unrestricted Bash/Python tools for malicious purposes beyond image generation.
SKILL.md
allowed-tools: Not specifiedWhat is in the package
| File | Kind | Size |
|---|---|---|
SKILL.md | doc | 1.7 KB |
scripts/generate.js | script | 5.6 KB |
External hosts referenced
daily-cloudcode-pa.sandbox.googleapis.com
Hostnames that appear in the skill's files. Referencing a host is not a finding; compare against the findings above and the skill's stated purpose.
Environment variables referenced
PROFILE_PATH
Package contents recorded 2026-10-22. The scan above ran against the version current at scan time; the two can differ if the skill has since been updated.
What we checked
- Static analyzer. Pattern rules over SKILL.md and bundled files: known-bad commands, exfiltration URLs, secrets, obfuscation.
- Behavioral analyzer. Checks what the skill asks the agent to do against what its description claims it does.
- Llm analyzer. A language model reads the skill and reports intent-level issues the rules cannot express.
- Trigger analyzer. Looks for instructions that fire on conditions the user did not ask for.
- Meta analyzer. A second model pass that validates, deduplicates and re-scores the other analyzers' findings.
What we did not check
- Runtime behaviour. The skill was not executed; findings come from reading its files.
- Anything fetched at install or run time, including remote scripts, package dependencies, and URLs the skill tells the agent to open.
- Content the analyzers cannot read: images, binaries, and some non-English text.
- Known scanner bypass classes. See the methodology and limitations page.
Sources
- Repository: https://clawhub.ai/skills/antigravity-image-gen
- Raw scan result: antigravity-image-generator-scan.json
- Badge for your README:
- Request a re-scan