endman100/skill-router-nexus/autoscirub-verify

Trust tier 3
  • community
Source check pending

Checking that the SKILL.md file is still in the repository.

Verify a scientific report and its code, results, tables, and figures criterion by criterion against an existing `.autoscirub/executable_rubric.json`, then produce concrete evidence gaps and optionally hand failed criteria to one bounded targeted revision round. Use only when an AutoSciRub rubric already exists or the user explicitly requests AutoSciRub verification. This scientific artifact check supplements rather than replaces general completion verification.

Install

npx vskill@latest install endman100/skill-router-nexus/autoscirub-verify

SKILL.md

AutoSciRub Verify

Use this verification-only entry when rubric induction is already complete. If no .autoscirub/executable_rubric.json exists, stop and route to the full autoscirub workflow in AI-Research instead of inventing criteria after execution.

Verify

  1. Read the executable rubric, current report, and supporting code, results, tables,

figures, logs, and analysis files.

  1. Check each criterion independently and preserve its stable identifier.
  2. Mark a criterion satisfied only when both the requested analysis and inspectable

task-generated evidence exist.

  1. Do not substitute literature claims for current-task evidence.
  2. Record evidence found, the exact gap, required corrective actions, and artifact paths.
  3. Write .autoscirub/verification_report.json or the requested round-specific file.
  4. Set all_satisfied true only when all required and high-priority criteria pass and

lower-priority exceptions are explicitly justified.

The report must include passed, failed, and highest_priority_gaps counts or lists.

Optional bounded revision

Run a revision only when the user or full AutoSciRub controller requested it and budget remains. Perform exactly one revision per invocation:

  1. Select failed criteria in priority order.
  2. Write a round-specific revision plan.
  3. Add missing analyses or comparisons, regenerate evidence, strengthen supported

explanations, or remove unsupported claims.

  1. Record changed artifacts and remaining gaps.
  2. Return to verification; a changed file is not proof that a criterion now passes.

Do not broaden the research objective beyond failed criteria. If a gap cannot be solved with visible data or tools, report the limitation instead of fabricating evidence.

Relationship to Router verification

This skill evaluates scientific adequacy against the executable rubric. Always run the Router's general verification-before-completion checks separately for commands, exit codes, tests, file presence, and the final completion claim.

Adapted from zjunlp/AutoSciRub, MIT licensed.

Security

  • Security scanPassed, 90 of 100Pattern checks on the skill's files.
  • Model reviewPassed

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

7-day trend score
5

Not refreshed yet.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/endman100/skill-router-nexus/autoscirub-verify/badge)](https://verified-skill.com/skills/endman100/skill-router-nexus/autoscirub-verify)

Something wrong with this skill or its page? Tell us.