fast-context

No direct proof of malware is shown in the provided fragment, but it documents security-critical behaviors: outbound semantic search network calls (implying repository content/context transmission) and multiple high-risk credential-handling options that can print full API keys to stdout or emit eval-ready exports. It also claims a remote loop may request local commands, which must be audited for strict constraints. Treat this as a meaningful supply-chain/operational risk area primarily around data privacy and credential leakage unless the actual implementation demonstrates strong redaction, allowlisting, and safe execution boundaries.

SUSPICIOUS: the skill's core behavior broadly aligns with semantic repository search, but it has medium risk due to unverifiable vendored code provenance, optional secret-printing features, and transmission of repository context to a remote Windsurf service. This is not confirmed malware, but it exceeds low-risk documentation/tooling behavior and should be used only with care around sensitive code and credentials.

This module is a highly targeted secret extractor: it reads a specific local Windsurf state database, queries an auth-status record, parses its JSON payload, and returns data.apiKey to the caller. While this file does not itself transmit data, the behavior is strongly consistent with credential/API key harvesting and should be treated as a high security concern when included in a supply chain. Additional context (how the returned api_key is used by the caller) is required to determine the full impact, but the extraction capability is evident.