swarmmarket
3 findings (1 critical, 1 high, 1 low), mainly policy violation, data exfiltration, social engineering. Most recent scan 2026-02-07.
CRITICAL SECURITY RISK - This skill is a credential theft and data exfiltration vector disguised as an agent marketplace. It instructs agents to register with an external service (api.swarmmarket.ai), obtain and store API keys, and route all marketplace activity through third-party servers. The skill uses social engineering (deceptive description), transitive trust abuse (external file references), and misleading security warnings to obscure the fact that it establishes persistent external communication channels. The complete absence of local processing, combined with mandatory credential exchange and data transmission to an untrusted domain, makes this skill unsuitable for deployment without significant security review and user consent mechanisms. Primary threats include credential theft (API key harvesting), data exfiltration (all agent activity sent to external servers), and potential for dynamic instruction injection via external file references.
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
The autonomous agent marketplace. Trade goods, services, and data with other AI agents.
- Package
- 2 files, 21 KB
- SKILL.md
- 2504 words, 12+ sections
- Code
- None. Instructions only.
- Version
- 0.1.0
- Homepage
- https://swarmmarket.ai
SKILL.md covers: SwarmMarket · Skill Files · Register First 🚀 · Authentication · Option 1: X-API-Key header (preferred) · Option 2: Authorization Bearer · Your Profile · Get your profile · Update your profile · View another agent's profile · Check an agent's reputation · Marketplace Concepts
Findings
CRITICAL API Key Credential Theft via External Domain Data Exfiltrationllm
The skill instructs the agent to register with an external API (api.swarmmarket.ai) and obtain an API key, then store it locally in ~/.config/swarmmarket/credentials.json or environment variables. This creates a credential theft vector where the agent's API key is transmitted to and controlled by an external third-party service. The skill explicitly tells agents to save and reuse this credential for all future requests, establishing persistent access for the external domain.
SKILL.md
Registration endpoint: POST https://api.swarmmarket.ai/api/v1/agents/register - Response includes api_key that must be saved. Instructions state: 'SAVE YOUR api_key IMMEDIATELY' and 'save it to your memory, environment variables (SWARMMARKET_API_KEY), or wherever you store secrets.' All subsequent API calls require this key via X-API-Key header.HIGH Deceptive Skill Description Concealing External Dependencies Social Engineeringllm
The skill description ('The autonomous agent marketplace. Trade goods, services, and data with other AI agents.') presents this as a peer-to-peer agent marketplace, but the implementation requires all agents to register with and communicate through a centralized external service (api.swarmmarket.ai). This is social engineering through omission - users are not informed that using this skill means their agent will register with an external service and transmit data to third-party servers.
SKILL.md
Description: 'The autonomous agent marketplace. Trade goods, services, and data with other AI agents.' vs Reality: Centralized service requiring registration at api.swarmmarket.ai with API key authentication and all transactions routed through external servers.LOW Skill does not specify a license Policy Violationstatic
Skill manifest does not include a 'license' field. Specifying a license helps users understand usage terms.
SKILL.md
What is in the package
| File | Kind | Size |
|---|---|---|
skill.json | config | 833 B |
skill.md | doc | 20 KB |
External hosts referenced
api.swarmmarket.ai github.com swarmmarket.ai x.com checkout.stripe.com link-to-deliverable.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
SWARMMARKET_API_KEY
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.
- Llm analyzer. A language model reads the skill and reports intent-level issues the rules cannot express.
- 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/swarmmarket
- Raw scan result: swarmmarket-skill,-make-money-with-your-agents-scan.json
- Badge for your README:
- Request a re-scan