Skip to content

feat: update research skills part 9#55

Merged
markusha77 merged 2 commits into
ChatAndBuild:mainfrom
sixiang-svg:update-research-part-9
Mar 25, 2026
Merged

feat: update research skills part 9#55
markusha77 merged 2 commits into
ChatAndBuild:mainfrom
sixiang-svg:update-research-part-9

Conversation

@sixiang-svg
Copy link
Copy Markdown
Contributor

New Skill Contribution

Skill Info

  • Skill ID:
  • Category:
  • Description:

Checklist

  • SKILL.md has valid YAML frontmatter with all required fields
  • Skill ID matches the directory name
  • Category is one of: productivity, development, communication, writing, research, other
  • Instructions are clear and self-contained
  • Instructions are under 4000 tokens
  • No external URLs referenced in instructions
  • Tested the skill with an AI agent and it produces good results
  • Added usage examples in the frontmatter

Testing

Describe how you tested this skill:

Notes

Any additional context or notes for reviewers.

@greptile-apps
Copy link
Copy Markdown
Contributor

greptile-apps Bot commented Mar 18, 2026

Greptile Summary

This PR adds 15 new/expanded research skills (tides, torchdrug, tracking-regression-tests, transit-least-squares, units, update-flaky-tests, usgs-data-download, uspto-database, uv-global, venue-templates, wandb-experiment-logger, weather-query, wiring, zemax-optical-designer, zinc-database) and intentionally deletes uniprot-database to resolve the H1/frontmatter mismatch flagged in a prior review. Most skills are well-structured upgrades from stubs, but two skills have meaningful quality or runtime issues.

  • update-flaky-tests: References four companion files (flaky-tests-schema.json, flaky-tests-sample.json, flaky-tests-archived-sample.json, status-output-sample.md) at the /.claude/skills/update-flaky-tests/ runtime path. None of these files exist in the skill's directory, meaning the schema-based initialization step will fail for any user who invokes this skill.
  • wandb-experiment-logger: The Instructions, When to Use, and Prerequisites sections are auto-generated boilerplate (e.g., "Provides step-by-step guidance for wandb experiment logger"). The skill provides no actionable W&B operational guidance and would not help an AI agent perform actual Weights & Biases tasks.
  • tracking-regression-tests: Alone among all 15 skills in this PR, it has no top-level H1 heading — the content opens directly with ## Overview.
  • wiring: The H1 reads # Admin Developer Guide rather than # Wiring, creating a confusing disconnect between the skill entry point and its document title.
  • torchdrug: Frontmatter uses name: Torchdrug; the H1 and official library branding both use TorchDrug.
  • Category misclassifications for tracking-regression-tests, update-flaky-tests, uv-global, and wiring were already flagged in a prior review round and remain unresolved.

Confidence Score: 3/5

  • Safe to merge for most skills, but update-flaky-tests will break at runtime due to missing schema files, and wandb-experiment-logger provides no actionable content.
  • The majority of skills (tides, transit-least-squares, units, usgs-data-download, uspto-database, venue-templates, weather-query, zemax-optical-designer, zinc-database, torchdrug) are high quality and ready to merge. However, update-flaky-tests has a concrete runtime failure path (missing companion files), and wandb-experiment-logger is essentially a placeholder that will not help users. Combined with four unresolved category misclassifications carried over from a previous review round, the PR warrants a mid-range confidence score.
  • skills/Research/update-flaky-tests/SKILL.md (missing companion schema/sample files), skills/Research/wandb-experiment-logger/SKILL.md (boilerplate placeholder instructions)

Important Files Changed

Filename Overview
skills/Research/tides/SKILL.md Expanded from a stub to a fully detailed skill with proper Instruction, When to Use, and Output sections; correctly categorized as Research.
skills/Research/torchdrug/SKILL.md Comprehensive new skill for PyTorch drug discovery workflows; frontmatter name "Torchdrug" should be "TorchDrug" to match the H1 and official library branding.
skills/Research/tracking-regression-tests/SKILL.md New regression-test tracking skill; missing a top-level H1 heading (unique among all skills in this PR) and miscategorized as Research instead of development.
skills/Research/uniprot-database/SKILL.md Intentionally deleted in this PR (commit: "remove skill: body dismatches frontmatter(bisoservices duplicate)") to address the H1/name mismatch flagged in a previous review.
skills/Research/update-flaky-tests/SKILL.md Detailed flaky-test tracker skill that references four companion files (schema, samples) at runtime, none of which exist in the skill directory — initialization will fail without them. Also miscategorized as Research.
skills/Research/uv-global/SKILL.md Clear Python tool management skill for uv; content is well-written but miscategorized as Research instead of development.
skills/Research/wandb-experiment-logger/SKILL.md The Instructions, When to Use, and Prerequisites sections contain auto-generated boilerplate that provides no actionable W&B guidance; an AI agent using this skill would receive no useful operational direction.
skills/Research/wiring/SKILL.md Detailed MobX/Webiny wiring guide; H1 reads "Admin Developer Guide" instead of "Wiring", and skill is miscategorized as Research instead of development.
skills/Research/zinc-database/SKILL.md Comprehensive ZINC cheminformatics skill with best practices, query optimization guidance, and appropriate data reliability disclaimers.

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart TD
    PR["PR #55 — 16 files changed"]

    PR --> DEL["🗑 uniprot-database/SKILL.md\nDeleted (H1 mismatch fix)"]
    PR --> GOOD["✅ Well-formed skills (9)"]
    PR --> WARN["⚠️ Minor issues (4)"]
    PR --> BAD["❌ Needs rework (2)"]

    GOOD --> G1["tides"]
    GOOD --> G2["transit-least-squares"]
    GOOD --> G3["units"]
    GOOD --> G4["usgs-data-download"]
    GOOD --> G5["uspto-database"]
    GOOD --> G6["venue-templates"]
    GOOD --> G7["weather-query"]
    GOOD --> G8["zemax-optical-designer"]
    GOOD --> G9["zinc-database"]

    WARN --> W1["torchdrug\nname: 'Torchdrug' → 'TorchDrug'"]
    WARN --> W2["tracking-regression-tests\nMissing H1 + wrong category"]
    WARN --> W3["uv-global\nWrong category: Research → development"]
    WARN --> W4["wiring\nH1 mismatch + wrong category"]

    BAD --> B1["update-flaky-tests\nMissing companion files\nSchema/samples not in repo"]
    BAD --> B2["wandb-experiment-logger\nBoilerplate placeholder instructions\nNo actionable W&B guidance"]
Loading

Comments Outside Diff (2)

  1. skills/Research/update-flaky-tests/SKILL.md, line 41 (link)

    P1 Missing companion schema files referenced at runtime

    Step 1 instructs the skill to initialize the database using /.claude/skills/update-flaky-tests/flaky-tests-schema.json, and the Reference Files section (lines 147–150) lists four more companion files that are expected at that path:

    • flaky-tests-schema.json (used for initialization — critical)
    • flaky-tests-sample.json
    • flaky-tests-archived-sample.json
    • status-output-sample.md

    None of these files exist in the skill's directory (skills/Research/update-flaky-tests/). When deployed, Claude will look for them at /.claude/skills/update-flaky-tests/<filename> — a path that is populated from the repository directory. Since the files are absent, the schema-based initialization step will fail for any new user of this skill.

    These companion files must be added alongside SKILL.md in the same directory for the skill to function correctly.

  2. skills/Research/tracking-regression-tests/SKILL.md, line 12-13 (link)

    P2 Missing H1 heading

    Every other skill updated in this PR opens its content with a top-level # <Skill Name> heading (e.g., # Tides, # Transit Least Squares, # USGS Data Download). This file jumps straight to ## Overview without an H1, breaking the consistent heading hierarchy and making the document harder to parse.

Last reviewed commit: "remove skill: body d..."

This skill enables Claude to track and run regression tests, ensuring new changes don't break existing functionality. It is triggered when the user asks to "track regression", "run regression tests", or uses the shortcut "reg". The skill helps in maintaining code stability by identifying critical tests, automating their execution, and analyzing the impact of changes. It also provides insights into test history and identifies flaky tests. The skill uses the `regression-test-tracker` plugin.
description: Track and run regression tests using the regression-test-tracker to ensure code stability and identify failures.
category: Research
requires: []
Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

P2 Incorrect category for development/testing skill

This skill covers software regression testing workflows, which is clearly a development category skill — not research. The same misclassification applies to several other skills in this PR:

  • skills/Research/tracking-regression-tests/SKILL.md:6category: Research → should be development
  • skills/Research/update-flaky-tests/SKILL.md:6category: Research → should be development
  • skills/Research/uv-global/SKILL.md:6category: Research → should be development
  • skills/Research/wiring/SKILL.md:6category: Research → should be development

All four skills describe developer tooling, test management, package management, or frontend architecture patterns — none of which belong in the research category. Using the wrong category will cause these skills to surface incorrectly in category-filtered searches.

Suggested change
requires: []
category: development

Comment thread skills/Research/uniprot-database/SKILL.md Outdated
@markusha77 markusha77 merged commit c5d2074 into ChatAndBuild:main Mar 25, 2026
3 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants