Prompt forest
Skill herehere14/Cascade-An-AI-Agent-World-Simulator/extensions/prompt-forest/skills/prompt-forest
Financial and Economic world prediction simulation. Derived from individual Agent emotions and behaviours using our heart engine, and ripple system can cause different butterfly effects on the simulated world, and make crucial decisions that shapes the macro world, economy, society, company. Achieved over 90% accuracy for covid 19 and 2008 f crises
npx -y skills add herehere14/Cascade-An-AI-Agent-World-Simulator --skill prompt-forestAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
2 things to look at
- no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
- 3 stars3 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Adaptive prompt routing for planning, verification, critique, code review, and confidence-qualified responses via the Prompt Forest plugin tools.
SKILL.md
1.5 KB, as published. Nobody here has run it
Prompt Forest
Use prompt_forest_assist when the user asks for work that benefits from structured branch routing rather than a single raw draft.
Good triggers:
- planning, rollout, roadmap, milestones, owners, risks, tradeoffs
- verification, auditing, contradiction checks, constraint satisfaction, evidence checks
- code review, bug triage, refactor planning, algorithm comparison
- critique, compare-and-choose, structured brainstorming, confidence-qualified answers
Avoid it for:
- trivial chit-chat
- one-line factual replies when no structure or validation is needed
- cases where the user only wants you to restate or lightly rewrite their text
How to call prompt_forest_assist:
- Put the full user request into
task. - Use
taskTypeonly when obvious; otherwise keepauto. - Pass
requiredChecksfor hard requirements such asconfidence,owners,risks, orcitations. - Pass
expectedKeywordsonly for the most important content anchors. - Put only high-signal context into
metadataorcontextSeed.
After the tool returns:
- answer naturally for the user instead of dumping raw metadata
- use the Prompt Forest metadata when it helps, especially
selected_branch,reward,confidence, and the trace - if the user explicitly accepts, rejects, or corrects the result, call
prompt_forest_feedback