duki

What an agent can do with the llm-output-guard corpus

duki-core/llm-output-guard: Lightweight, vendor-agnostic library for detecting system prompt leaks and gating unauthorized action markers in LLM output

• Look up how llm-output-guard implements ai-safety, llm-security, python, with answers cited to file and line.
• Reuse a worked example: the commit “Add StreamGuard: chunk-safe marker filtering for streamed output” (5 files · Python).
• Answer from llm-output-guard's real Python source instead of memory (AI, prompts and agents).

• Where does llm-output-guard handle “Add StreamGuard: chunk-safe marker filtering for streamed output”?
• How does llm-output-guard implement ai-safety?
• What does tests do in llm-output-guard?

Search the GitHub repository duki-core/llm-output-guard with XERJ (https://hub.xerj.org/corpus/duki-core--llm-output-guard).
Run these in a shell, in order, and check each result before the next:
1. xerj --help  (success: it prints usage). If xerj is missing: curl -fsSL https://xerj.org/get | sh
2. git clone https://github.com/duki-core/llm-output-guard llm-output-guard && git -C llm-output-guard checkout 63dcb657a71d7ca8b86fbcb17a4469b111153428
3. Start a local node and leave it running: xerj --insecure --data-dir ./xerj-data
4. xerj autoindex ./llm-output-guard --url http://127.0.0.1:9200 --prefix llm-output-guard --progress plain
5. xerj autoindex map --url http://127.0.0.1:9200
Then answer my question from the indexed files and cite the ax_path of every hit. Query syntax: https://xerj.org/llms.txt
Licence: MIT. Keep the copyright notice and credit the authors.

corpus duki-core--llm-output-guard · AI, prompts and agents
Related corpora: bandit-plugins, garak, Actenon/actenon-scan, saintparish4/openscript
AI search and retrieval with XERJ · Case study: reference coding for agents