gonzalezpazmonica/model-upgrade-audit
.claude/skills/model-upgrade-audit/SKILL.md
Audit workspace components for prompt debt when a new model is available. Detect workarounds, propose simplifications, compare with evals.
$ install --global
skillsauthnpx skillsauth add gonzalezpazmonica/pm-workspace model-upgrade-auditInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
Security Scan Results
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
3
Scanners Passed
9
Scanners in report
Clean
TrivyContainer and dependency vulnerability scanner
95%
Clean
SemgrepStatic code analysis for vulnerabilities
95%
Clean
mcp-scan (Snyk)Model Context Protocol security validation
95%
Skipped
Snyk (dep)Open source security scanning
50%
Skipped
Socket.devSupply chain security analysis
50%
Skipped
VirusTotalMulti-engine malware detection
50%
Skipped
CrowdStrikeAdvanced threat intelligence
50%
Skipped
OSV-ScannerOpen Source Vulnerability database check
50%
Skipped
OWASP Dep-Check
50%
Last scanned: Apr 24, 2026, 8:50 PM1.7s1 file scanned
SKILL.md
- name:
- model-upgrade-audit
- description:
- Audit workspace components for prompt debt when a new model is available. Detect workarounds, propose simplifications, compare with evals.
- summary:
- |
- Ejecutar al cambiar de modelo (ej:
- Sonnet 4 a 4.5).
- category:
- governance
- tags:
- [model, upgrade, prompt-debt, simplification, audit]
- priority:
- medium
- disable-model-invocation:
- false
- user-invocable:
- true
- allowed-tools:
- [Read, Write, Glob, Grep, Bash, Task]
Model Upgrade Audit
Analyzes agents, skills, and prompts for workarounds that newer models handle natively. Proposes token-saving simplifications backed by evidence.
Decision Checklist
- Is a new model available in config? If NO -> abort, nothing to audit.
- Scope defined? (full | list | changed-since) If NO -> default to full.
- Are evals available for target components? If NO -> warn, audit without comparison.
- Is this a major model jump (e.g., Sonnet -> Opus)? If YES -> recommend full scope.
- Has previous audit been run? If YES -> show delta since last audit.
Parameters
| Param | Required | Default | Description |
|---|---|---|---|
| scope | No | full | full, agents, skills, rules, or comma-separated list |
| model_new | No | (auto-detect) | Model to audit against |
| changed_since | No | - | Only audit components changed since Era N |
Workaround Patterns
| Pattern | Signal | Severity | |---|---|---| | Emphatic repetition | Same instruction >= 2x | Medium | | Negative overload | >3 "don't/never/avoid" per prompt | Low | | Compensatory few-shot | Basic capability examples | Medium | | Defensive parsing | Regex fallback for bad output | High | | Coded retries | Retry loops for model failure | High | | Bloated prompt | >2000 tokens procedural | Medium |
Execution Flow
1. Inventory: glob agents/*.md + skills/*/SKILL.md + rules/domain/*.md
2. For each component:
a. Count tokens (wc -w * 1.3)
b. Scan for workaround patterns (regex)
c. Classify: simplifiable | no_change | review_needed
3. For simplifiable components:
a. Propose simplified version
b. Estimate token reduction
4. Generate YAML report -> output/model-audit/
5. Summary in chat (output-first pattern)
Output
Report: output/model-audit/{date}-audit.yaml
Summary:
Model Audit: {model_old} -> {model_new}
Components: {N} audited | {N} simplifiable | {N} no change
Token savings: ~{N} tokens/session ({pct}% reduction)
Risk: {N} low | {N} medium | {N} high
Application Flow
APPLY(risk: low) -> auto-apply with backupREVIEW(risk: medium) -> Draft PR for human reviewSKIP(risk: high) -> log for manual review
Integration
- Consumes feasibility-probe historical data for comparison
- Feeds into
/agent-efficiencyfor token tracking - Triggers after model config change in pm-config.md
Related Skills
gonzalezpazmonica/skill-creator
testing
VerifiedTrustedCommunity
Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.
40SKILL.mdUpdated May 24, 2026
gonzalezpazmonica/mcp-builder
tools
VerifiedTrustedCommunity
Guide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK).
40SKILL.mdUpdated May 24, 2026
gonzalezpazmonica/wellbeing-guardian
tools
VerifiedTrustedCommunity
Sistema proactivo de bienestar individual
33SKILL.mdUpdated Apr 17, 2026
gonzalezpazmonica/web-research
development
VerifiedTrustedCommunity
Search the web to resolve context gaps — documentation, versions, CVEs, best practices. Auto-starts SearxNG Docker if available, falls back to WebSearch.
33SKILL.mdUpdated Apr 17, 2026