frontier-gradient-step
暂无简介
- License 许可
- MIT
- 最近更新
- 2026-06-13
- Stars
- ★ 3
- Forks
- 1
01 前言 Front Matter
- id
- frontier-gradient-step
- title
- Gradient Step
03 SKILL.md 正文
Gradient Step
Update the accumulated knowledge model and decide whether to continue or converge.
Convergence Rule
if (currentDelta < threshold AND priorDelta < threshold):
decision = "CONVERGED" # 2 consecutive low-delta epochs
else:
decision = "CONTINUE"
The threshold is the convergenceThreshold input (default: 0.15).
Two consecutive low-delta epochs are required to avoid premature convergence on a single unproductive epoch.
Process
- Read selection results (insight delta, merged knowledge)
- Read prior
research-state.jsonandresearch-ledger.jsonl - Update knowledge model:
- Add new claims with source epoch and beam
- Add new connections between claims
- Add new dead ends
- Update convergence history
- Evaluate convergence:
- Get current insight delta from selection
- Get prior insight delta from ledger
- Count consecutive low-delta epochs
- Apply convergence rule
- Append metrics to ledger
- If CONTINUE, provide guidance for next epoch (which frontier areas need attention)
- If CONVERGED, summarize the final knowledge state
Outputs
gradient-step.json— convergence decision with rationale- Updated
research-state.json— accumulated knowledge model - Appended line to
research-ledger.jsonl— epoch metrics
Quality Criteria
- Convergence rule is applied strictly (not subjective judgment)
- Knowledge model is cumulative (prior claims preserved, not overwritten)
- Ledger is append-only (prior lines not modified)
- Next-epoch guidance is specific and actionable (for CONTINUE decisions)
- Final summary captures the full research journey (for CONVERGED decisions)
04 同仓库更多 Skills
hello-checker
Use this skill when asked to "verify", "check", or "validate" a file's existence and content. It provides a reliable workflow for confirming files meet their spec.
ask-skill
暂无简介
html-review-artifact
Framework-wide guidance for crafting polished human-review artifacts as markdown or HTML
converge-planning
Design a Converge playbook end-to-end. Use when the user says "plan a project", "design a playbook", "decompose into tasks", "scaffold a Converge workflow", "onboard to Converge", or asks how to write TASK.md / playbook.yml. Two entry points: plan a new playbook or update an existing one. Hand off to `/converge-control` after producing the playbook structure.
converge-control
Operate a Converge playbook end-to-end — start runs, monitor progress, narrow failures with `--select`, inspect runtime state, validate definitions, surgically reset subtrees, and stop stuck or orphaned runs. Use this skill whenever the user says "run my playbook", "execute this", "monitor the run", "converge failed", "the run is stuck", "retry failures", "babysit this playbook", "what's the status?", "show me the gantt", or asks how to use `converge run` / `--select` / `inspect` / `doctor` / `clean` / `stop`. Also use it for `HTTP 401` / `Invalid API key` failures on the first task — those are environment-not-playbook problems — and for any post-run cleanup work. Hand off to `converge-planning` if the user wants to design or restructure the playbook itself.
repair-taskts
Fix task.ts definition issues — Seed syntax, missing .build(), empty checks, import errors, TypeScript errors