simplify-jules
Skill65 starsChanged 7 months agoMITView on GitHub
On-demand config hygiene — analyzes cold-start bundle for waste, duplication, and stale content
In this repository19 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- agent-browserSkill
- buildSkill
- debugSkill
- financial-advisorSkill
- generate-image-openaiSkill
- good-morningSkill
- pdfSkill
- planeSkill
- reply-xSkill
- researchSkill
- search-historySkill
- send-emailSkill
- skill-creatorSkill
- stop-slopSkill
- thinkSkill
- watch-contactsSkill
- wrap-upSkill
- writeSkill
The file
We split this skill into the rules it gives an agent
7 rules. About 0 are common, 7 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈7
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 14
Read and count lines in every cold-start file:
- Distinctive: line 60
After reading all files, scan for:
- Distinctive: line 144
Write the full report to:
- Distinctive: line 150
Use today's date. If a file with today's date already exists, append a new timestamped section rather than overwriting.
- Distinctive: line 164
Never auto-edit any file: Read-only analysis only.
- Distinctive: line 165
All proposals require [Your Name]'s approval: before any changes are made.
- Distinctive: line 167
Use the same proposal pipeline as the retro process: present, discuss, [Your Name] approves, then implement in a separate session.
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
/simplify-jules
On-demand config hygiene tool. Analyzes the full cold-start file bundle for waste, duplication, stale content, and consolidation opportunities. All proposals require [Your Name]'s approval — never auto-edit any file.
Phase 1 — INVENTORY
Read and count lines in every cold-start file:
~/.claude/CLAUDE.mdand all@-imported files it referencesActive-Work/CLAUDE.mdand all@-imported files it references- All
.claude/rules/*.mdfiles - All
@-tagged profile files (Profiles/*.md)
For each file, record: file path, line count.
Report:
Cold-Start Budget: X lines
Files scanned: N
Largest files: [top 5 by line count]Phase 2 — SECTION ANALYSIS
For each file, read every section (heading or logical block) and classify it:
| Classification | Definition |
| ACTIONABLE | Directly changes agent behavior — keep as-is |
| DOCUMENTATION | Describes or explains but doesn't change behavior — propose move to Documents/ |
| REDUNDANT | Already said elsewhere — propose cut, cite exact duplicate (file:line) |
| VAGUE | Too abstract to change behavior — propose cut or rewrite |
Output one table per file:
### ~/.claude/CLAUDE.md
| Section | Lines | Classification | Notes |
|---------|-------|----------------|-------|
| User Preferences | 1-3 | DOCUMENTATION | Says "see project CLAUDE.md" |
...Cite line numbers for every section.
Phase 3 — CROSS-FILE DEDUP
After reading all files, scan for:
- Duplicate instructions — same rule stated in multiple files. List both locations.
- Conflicting instructions — contradictory rules. Flag both locations, describe the conflict.
- Consolidation opportunities — related content spread across files that could merge without loss.
Format:
### Duplicates
- "Stage specific files for commits" — rules/safety.md:12 AND CLAUDE.md:47
### Conflicts
- "Use Opus for complex analysis" (CLAUDE.md:89) vs "Default to Sonnet" (rules/models.md:5) — unclear precedence
### Consolidation
- Agent delegation rules split across CLAUDE.md:82-95 and rules/request-routing.md:44-60 — could mergePhase 4 — STALENESS CHECK
Flag the following:
- Past dates — any reference to specific dates that have already passed (e.g., "Last updated: 2025-03-01")
- Financial data older than 30 days — check
personal-finance-health.mdand any embedded financial data; flag if last review date is >30 days ago - Skills not invoked in 30+ days — check
Documents/Field-Notes/Logs/for skill invocation history. Cross-reference against all@-tagged skills and skill references in CLAUDE.md. Flag skills with no invocation record in the last 30 days.
Format:
### Stale Dates
- rules/personal-finance-health.md:5 — "Last updated: 2026-03-27" (N days ago)
### Stale Financial Data
- rules/personal-finance-health.md — last review 2026-03-27, N days ago. Flag for /financial-advisor refresh.
### Underused Skills (30+ days no invocation)
- /good-morning-demo — no log entry found
- /write-demo — no log entry foundPhase 5 — PROPOSALS
Generate an IVE-scored proposal list:
IVE = Impact × Velocity × Efficiency (each 1–3, product 1–27)
| Dimension | 1 | 2 | 3 |
| Impact | Minor improvement | Moderate improvement | Significant improvement |
| Velocity | Slow to implement | Medium effort | Quick win |
| Efficiency | Few lines saved | Moderate savings | Many lines saved |
Sort proposals descending by IVE score.
For each proposal:
## [IVE: 18] Cut redundant safety rules from CLAUDE.md
**What:** Remove lines 44-52 in Active-Work/CLAUDE.md (git staging rules)
**Why:** REDUNDANT — identical content at rules/safety.md:10-18
**Lines saved:** 9
**Risk:** Low — content preserved at canonical location
**Cite:** Active-Work/CLAUDE.md:44-52 duplicates .claude/rules/safety.md:10-18End with:
### Projected Budget
Before: X lines
After (if all proposals accepted): Y lines
Savings: Z lines (N%)Phase 6 — SAVE REPORT
Write the full report to:
Documents/Field-Notes/Research/YYYY-MM-DD-Simplification-Scan.mdUse today's date. If a file with today's date already exists, append a new timestamped section rather than overwriting.
Report header format:
# Simplification Scan — YYYY-MM-DD HH:MM
Cold-Start Budget: X lines | Files: N | Run by: /simplify-julesHard Rules
- Never auto-edit any file. Read-only analysis only.
- All proposals require [Your Name]'s approval before any changes are made.
- Cite line numbers everywhere — every classification, every duplicate, every proposal.
- Use the same proposal pipeline as the retro process: present, discuss, [Your Name] approves, then implement in a separate session.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

