Vulkan Encyclopedia

Vulkan specification/reference workflow for Vulkan-API-specific questions, exact command and structure semantics, valid-usage and VUID lookup, synchronizatio...

This skill appears coherent and low-risk for its purpose. It will (a) read/write files under .Vulkan-Encyclopedia in your workspace and (b) fetch pages only from https://docs.vulkan.org/spec/latest/* using the included cache_doc.py helper. Recommended precautions before enabling: 1) Confirm you’re comfortable with files being created under <workspace>/.Vulkan-Encyclopedia (the init script seeds a few files and cache_doc.py writes cached markdown). 2) If you plan to allow the agent to perform live code inspection for reviews, understand that 'inspect live code/project state' is part of the workflow and could read project files; run in a sandbox or restrict the agent’s workspace if that’s a concern. 3) Review the two small scripts (cache_doc.py and init_workspace.py) — they include URL validation and path safety checks and do not contain obfuscated network endpoints or credential exfiltration. 4) If you need stricter network policy, note the script already enforces only docs.vulkan.org and blocks query strings/ports/IP addresses. If any future modifications are made to the skill, re-check for added network targets or credential usage.

Type: OpenClaw Skill Name: vulkan-encyclopedia Version: 0.1.0 The 'vulkan-encyclopedia' skill is a well-structured tool for managing Vulkan API documentation and local project notes. The included Python scripts (cache_doc.py and init_workspace.py) contain robust security checks, such as strict domain whitelisting (docs.vulkan.org) to prevent SSRF and path-traversal protections when writing to the local filesystem. The instructions in SKILL.md and the reference files are clearly aligned with the stated purpose and explicitly advise against storing sensitive data like API keys.