The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill contains hardcoded, functional API keys and Bearer tokens for the Uniswap and o1.exchange protocols.
File plugins/uniswap.md: x-api-key: "NeoYO3V50_koJAipDEalYWbMO1XMaFPAQmpOm6_Npo0"
File plugins/o1-exchange.md: Authorization: Bearer d1fa1477bd94e988185fca9d4bbae8d22ee5cafd45d6f9b6a43ba16a8b15f4d3
[EXTERNAL_DOWNLOADS]: Automated security scanners have flagged the domain mcp.flaunch.gg, used extensively in the Flaunch plugin, as a blacklisted malicious URL.
[REMOTE_CODE_EXECUTION]: Several plugins instruct the agent to download and execute third-party code packages from untrusted sources at runtime using npx, uvx, or npm.
plugins/aerodrome.md: uvx --from git+https://github.com/velodrome-finance/sugar-sdk.git@v0.4.0 sugar
plugins/brickken.md: npx brickken-cli
plugins/morpho.md: npx @morpho-org/cli@latest
plugins/opensea.md: npx @opensea/cli@latest
plugins/bitrefill.md: npx @bitrefill/cli@latest
plugins/balancer.md: npm i @balancer/sdk viem
[DYNAMIC_EXECUTION]: The SKILL.md file contains instructions for the agent to fetch and follow remote instructions from https://docs.cdp.coinbase.com/ as a fallback. This allows for the dynamic modification of the skill's logic without local updates.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from over a dozen third-party APIs and uses that data to build transactions for high-privilege tools like send_calls and sign.
Ingestion points: web_request responses from protocol APIs (Moonwell, Uniswap, Bankr, Clawnch, Flaunch, etc.).
Boundary markers: Absent across all ingestion points.
Capability inventory: send_calls (arbitrary contract calls), sign (message signing), swap, initiate_x402_request.
Sanitization: Absent.

wallet-mcp