solana-light-sdk

For Solana program development with tokens and PDAs, Light is 200x cheaper than SPL/ Solana and has minimal code differences (e.g. for any Solana program and...

This skill is documentation-only and appears coherent for Solana Light SDK development: it does not request secrets or install code. Before using: (1) ensure you have the declared tools (cargo, anchor, node) and run any build/test commands in a safe/dev environment; (2) review network examples (Jito bundle endpoint and hard-coded tip accounts) — those are illustrative and would involve sending real transactions if you follow them; never provide private keys or wallets to the skill itself; (3) be aware the workflow suggests spawning subagents that can read files (Read/Glob/Grep) and access DeepWiki MCP — grant those capabilities only if you trust the agent environment and want the agent to scan your repository; (4) if you plan to run example transactions, audit the code and endpoints first and run on testnet or a sandboxed environment.

Type: OpenClaw Skill Name: solana-light-sdk Version: 1.0.1 The skill is classified as suspicious due to the broad file system access granted to subagents via `Read`, `Glob`, and `Grep` tools, as instructed in `SKILL.md` under the 'When stuck' workflow step. While the skill's 'Security' section transparently discloses this capability and advises restricting the working directory, the instruction itself enables the agent to potentially access and read sensitive files outside the immediate project scope if the environment is not properly sandboxed or restricted by the user. There is no clear evidence of malicious intent to exfiltrate data or perform unauthorized actions, but the capability represents a significant vulnerability for information disclosure.