AI/ML · Advanced pack
Hallucination Regression Hunter
Build eval cases that expose fabricated citations, brittle reasoning chains, and ungrounded tool usage before they hit real workflows.
CodexClaude CodeKimi Codeorchestrator-mcp
Best use case
Use Hallucination Regression Hunter when you need to build eval cases that expose fabricated citations, brittle reasoning chains, and ungrounded tool usage before they hit real workflows, especially when the work is driven by hallucination and grounding.
Trigger signals
hallucinationgroundingeval caseagent reliability
Validation hooks
verify_eval_coverageverify_text_unchanged
Install surface
Copy the exact command path you need.
Inspect
pip install "orchestrator-mcp[dashboard]"
orchestrator-mcp skills show hallucination-regression-hunter
Use
orchestrator-mcp skills export hallucination-regression-hunter --to ./skillforge-packs
# copy the exported pack into your preferred agent environment
Export
cp -R skills/hallucination-regression-hunter ./your-agent-skills/hallucination-regression-hunter
# or open skills/hallucination-regression-hunter/SKILL.md in a markdown-first client
File patterns
**/evals/****/tests/****/*.json**/*.md
Model preferences
Related skills
Adjacent packs to compose next.
Improve chunking, metadata, and ranking design so agent answers stay grounded under larger repositories and longer tasks.
CodexClaude Code
Allocate work across local, cloud, and premium models so teams maximize capability coverage per dollar and per latency budget.
CodexClaude Code
Reshape sprawling repositories and briefs into stable context lanes, memory checkpoints, and retrieval boundaries for long-horizon agent work.
CodexClaude Code