Code analysis
static source read inferred
Static code-analysis findings — hidden prompt content in shipped skill files, committed secrets, dynamic-exec sinks, and suspicious call-home endpoints — across the analyzed catalogue. Heuristic, pure, no code executed; every row deep-links to its source. Click a kind to filter.
50100 analyzed
8073 re-analysis due
988 not analyzable
2 not yet analyzed
4804 source gone
not analyzable
789 too large 199 no source
Running analyzer v33. The scanner changelog explains what each version detects and when it changed.
- hidden prompt 338
- committed secret 6180
- dynamic exec 11839
- obfuscation 3598
- suspicious endpoint 12356
- credential in log 713
- over-broad oauth scope 2550
- suspicious skill script 185
- bundled IDE extension 49
- skill file 187699
- HIGH dynamic exec HatriGt/agent-sandbox eval()
async eval(expression) { - HIGH dynamic exec HatriGt/agent-sandbox eval()
async eval(expression) { - HIGH committed secret HatriGt/agent-sandbox GitHub fine-grained PAT
github…(57 chars, redacted) - HIGH committed secret HatriGt/agent-sandbox Anthropic key
sk-ant…(39 chars, redacted) - MEDIUM suspicious endpoint HatriGt/agent-sandbox 169.254.169.254 (cloud metadata)
assert.equal(isProbeableUrl("https://169.254.169.254/latest/meta-data"), false); - MEDIUM suspicious endpoint mcp-hangar 169.254.169.254 (cloud metadata)
_, warnings = _run({"meta": {"mode": "remote", "endpoint": "http://169.254.169.254/latest/meta-data/"}}) - MEDIUM suspicious endpoint mcp-hangar 169.254.169.254 (cloud metadata)
"http://169.254.169.254/latest/meta-data/", - MEDIUM over-broad oauth scope mcp-hangar delete_repo
server.invoke_tool("delete_repo", {}) - MEDIUM suspicious endpoint mcp-hangar 169.254.169.254 (cloud metadata)
validate_no_ssrf("http://169.254.169.254/latest/meta-data/") - HIGH committed secret mcp-hangar Google API key
AIzaSy…(39 chars, redacted) - MEDIUM over-broad oauth scope mcp-hangar delete_repo
_server(L7Policy(tools=ToolRules(deny=("delete_*",)))).invoke_tool("delete_repo", {}) - MEDIUM over-broad oauth scope mcp-hangar delete_repo
action, _ = evaluate_tool("delete_repo", rules, ToolAction.DENY) - MEDIUM suspicious endpoint AlgoVaultLabs/crypto-quant-signal-mcp t.me
expect(html).toContain('https://t.me/algovaultofficialbot?start=auth_av_live_abc123def456789012345678'); - MEDIUM suspicious endpoint AlgoVaultLabs/crypto-quant-signal-mcp 8.8.8.8
await expect(resolveAndAssertEgress('https://8.8.8.8/x', { lookup: lk })) - MEDIUM suspicious endpoint AlgoVaultLabs/crypto-quant-signal-mcp 8.8.8.8
const pin = await resolveAndAssertEgress('https://8.8.8.8/x', { lookup });