Dsh Llm Headers
QiE2035/dsh-llm-headers · skills
The custom LLM API request for DeepSeek Harness is injected into HTTP Headers (typical use: rewriting user-agent, e.g. gateway requirement UA with specified brand).
★ 3GitHub stars
1Forks
2026-08-28Last updated
TypeScriptLanguage
MITLicense
Key features
- The LLM adapter for DeepSeek Harness (dsh-llm-deepseek, dsh-llm-pi-ai) sends a model request via global Fetch.
- This plugin is packaged on load with globalThhis. Fetch: Only for URL matching requests of urlPattens, the configured headers set() go up (covering the same name, including mandatory adaptation user-agent - i.e. deployment white label replacement scene), and the rest are faxed as it stands.
- The provider is irrelevant, and the one-time configuration works for all LLM projects.
- Configure not to writeThe plugin registers Config schema as llm-headers user configuration naming space, the Web UI's settings automatically render an edit sheet of that namespace and the next request becomes effective after saving (live heat load does not need to be restarted).
Requirements
- User-agent cover is deployment white mark replacement: mandatory decision-making in warehouse (dsh-llm/src/resolution.ts) does not allow inhibition of attribution per se, but allows replacement with deployment brand by status
- This plugin replaces the HTTP layer at an equal price only when you configure that head in a visible configuration in the headers.
Install command
dsh plugin --profile web add github:QiE2035/dsh-llm-headers