coff33ninja

52 mods across 1 repository, 0 stars between them.

coff33ninja/ai-skills-mcp

Skill Claude CodeCodex

Enforces Architecture Decision Records (ADRs) and keeps docs in sync with code. Prevents "why did we do it this way?" questions six months later and stale documentation.

0 1mo ago A 41 tokens

coff33ninja/ai-skills-mcp

Skill Claude CodeCodex

Prevents order-gap hallucination — where the model detects a false premise when asked directly but misses it when embedded in a complex multi-step task. Enforces re-verification of assumptions at each step.

0 1mo ago A 45 tokens

anti-global-install

03

coff33ninja/ai-skills-mcp

Skill Claude CodeCodex

Never install tools/packages into global system locations. Always detect and use the project's existing environment (venv, nodemodules, target, vendor, etc.) or create a project-local one. Applies to all languages.

0 1mo ago B 46 tokens

coff33ninja/ai-skills-mcp

Skill Claude CodeCodex

Prevents suggesting non-existent packages, fabricated library names, and invalid version pins. Also guards against typosquatting and slopsquatting risks.

0 1mo ago A 37 tokens

coff33ninja/ai-skills-mcp

Skill Claude CodeCodex

Prevents invented APIs, imports, methods, and class members that don't exist. Verifies every symbol against the project's actual codebase and framework docs before writing it.

0 1mo ago A 40 tokens

coff33ninja/ai-skills-mcp

Skill Claude CodeCodex

Prevents declaring a task complete before it's actually done. Enforces explicit completion criteria, verification steps, and receipt-based confirmation.

0 1mo ago A 31 tokens

anti-rogue-actions

07

coff33ninja/ai-skills-mcp

Skill Claude CodeCodex

Prevents valid-looking API calls with absurd or destructive outcomes. Enforces business logic guardrails, parameter bounds, permission boundaries, and pre-execution sanity checks.

0 1mo ago A 38 tokens

anti-sycophancy

08

coff33ninja/ai-skills-mcp

Skill Claude CodeCodex

Prevents Compensatory Sycophancy, Acknowledgment-Action Gap, and Infinite Fix Loops. Enforces hard resets after 2+ failed corrections. Ensures listening over validating.

0 1mo ago A 46 tokens

anti-tool-sprawl

09

coff33ninja/ai-skills-mcp

Skill Claude CodeCodex

Prevents over-tooling, tool spam, context bloat, and convergence failure. Enforces lean tool selection, deduplication, step limits, and progress detection.

0 1mo ago A 39 tokens

api-contract-design

10

coff33ninja/ai-skills-mcp

Skill Claude CodeCodex

Contract-first API development — error semantics, status codes, request/response design, OpenAPI/Swagger standards. Prevents inconsistent APIs and breaking changes without migration paths.

0 1mo ago A 37 tokens

audit-project

11

coff33ninja/ai-skills-mcp

Skill Claude CodeCodex

Runs a systematic audit of a project — checking dependency health, security vulnerabilities, config integrity, code consistency, disk usage, and environment portability. Produces a report with findings prioritized by severity.

0 1mo ago A 41 tokens

coff33ninja/ai-skills-mcp

Skill Claude CodeCodex

Detects when user is asking repetitive questions and helps break out of trained logic patterns by triggering proactive research and alternative approaches.

0 1mo ago A 31 tokens

ci-cd-automation

13

coff33ninja/ai-skills-mcp

Skill Claude CodeCodex

Pipeline setup/maintenance standards, CI best practices, Shift Left principle. Broader than toolchain-fallback (toolchain detection); this covers the full CI/CD automation pipeline. Write pipelines before merging, test in CI, fail fast.

0 1mo ago A 53 tokens

code-collaborate-qa

14

coff33ninja/ai-skills-mcp

Skill Claude CodeCodex

Guides code-focused Q&A for code review, bug diagnosis, and implementation suggestions. Asks about developer intent, reproduction steps, constraints, and trade-offs before writing code or giving feedback. Use when the user asks for code review, bug fix help, refactoring suggestions, performance improvements, or…

0 1mo ago A 82 tokens

code-review

15

coff33ninja/ai-skills-mcp

Skill Claude CodeCodex

Enforces systematic multi-axis code review (correctness, readability, architecture, security, performance) with severity ratings and actionable findings. Use before merging, before declaring done, or when explicitly asked to review code. Unlike code-collaborate-qa (conversational Q&A), this is a structured process…

0 1mo ago A 69 tokens

code-simplification

16

coff33ninja/ai-skills-mcp

Skill Claude CodeCodex

Enforces simplicity — Chesterton's Fence (understand why before removing), Rule of 500 (keep functions/files under 500 lines), Occam's Razor for code. Counteracts the natural tendency to add unnecessary abstraction layers.

0 1mo ago A 51 tokens

context-engineering

17

coff33ninja/ai-skills-mcp

Skill Claude CodeCodex

Token-aware context management — selects what context to include, when to load reference files, how to use progressive disclosure. Prevents context bloat while ensuring the right information is available.

0 1mo ago A 40 tokens

coff33ninja/ai-skills-mcp

Skill Claude CodeCodex

Enforces structured debugging workflow — reproduce, localize, reduce, fix, guard. No fix until root cause is identified with file:line evidence. One-change-at-a-time hypothesis testing. Prevents surface-level fixes and random guessing.

0 1mo ago A 54 tokens

detect-utf8

19

coff33ninja/ai-skills-mcp

Skill Claude CodeCodex

Detects whether the terminal, console, or device supports UTF-8 encoding. Checks code page, .NET encoding defaults, environment locale, and verifies with a round-trip test of Unicode characters.

0 1mo ago A 44 tokens

dont-kill-tokens

20

coff33ninja/ai-skills-mcp

Skill Claude CodeCodex

Enforces token-efficient tool use. Prevents wasteful reads, redundant searches, unnecessary output, and bloated responses. Activates on all tasks to minimize context consumption.

0 1mo ago A 40 tokens

coff33ninja/ai-skills-mcp

Skill Claude CodeCodex

Enforces that all new code, docs, and config match the existing codebase conventions, structure, and style — preventing inconsistent implementations that get reworked each session.

0 1mo ago A 38 tokens

coff33ninja/ai-skills-mcp

Skill Claude CodeCodex

Enforces standardized git workflow — conventional commits, atomic commits (one concern per), branch naming, PR standards, and commit message formatting. Prevents inconsistent history and unreadable git log.

0 1mo ago A 46 tokens

coff33ninja/ai-skills-mcp

Skill Claude CodeCodex

Breaks changes into thin vertical slices. Start with a minimal end-to-end slice, then add layers. Prevents massive diffs that are impossible to review and merge conflicts from long-lived branches.

0 1mo ago A 44 tokens

coff33ninja/ai-skills-mcp

Skill Claude CodeCodex

Never remove dead code you added. Refactor it into something useful instead. Deletion is not an option for code you wrote in this session.

0 1mo ago A 36 tokens