Code Amplifier

10 skills inclus

• 01

  Change Review (Standards and Spec)

  Fait relire par l'agent un diff depuis un point fixe sur deux axes séparés : conformité aux standards du dépôt et respect de la spec d'origine.+

  sf-change-review/SKILL.md, de mattpocock/skills, licence MIT.

  1---

  2name: sf-change-review

  3description: "Reviews the changes since a fixed point (commit, branch, tag or merge-base) along two separate…

  4---

  5

  6# Change Review (Standards and Spec)

  7

  8Two-axis review of the diff between `HEAD` and a fixed point the user supplies:

  9

  10- **Standards**: does the code conform to this repo's documented coding standards?

  11- **Spec**: does the code faithfully implement the originating issue / spec?

  12

  13Both axes run as **parallel sub-agents** so they don't pollute each other's context, then this skill aggregat…

  14

  15## Process

  16
• 02

  Requesting Code Review

  Fait envoyer par l'agent son travail terminé à un sous-agent relecteur avec un contexte précis, puis traiter les constats critiques, importants et mineurs.+

  sf-review-request/SKILL.md, de obra/superpowers, licence MIT.

  1---

  2name: sf-review-request

  3description: Dispatches a code-reviewer sub-agent with precisely crafted context (what was built, the require…

  4---

  5

  6# Requesting Code Review

  7

  8Send your finished work to a fresh reviewer sub-agent. The reviewer gets precisely crafted context, never you…

  9

  10## When to request

  11

  12Mandatory:

  13- After each task when executing a plan (`sf-plan-execution`).

  14- After completing a major feature.

  15- Before merging to main.

  16
• 03

  Receiving Code Review

  Fait traiter les retours de revue par l'agent avec rigueur : reformuler, vérifier dans le code, contester preuves à l'appui, corriger un point testé à la fois.+

  sf-review-feedback/SKILL.md, de obra/superpowers, licence MIT.

  1---

  2name: sf-review-feedback

  3description: Handles code review feedback with technical rigor instead of performative agreement or blind imp…

  4---

  5

  6# Receiving Code Review

  7

  8Code review requires technical evaluation, not emotional performance.

  9

  10**Core principle:** Verify before implementing. Ask before assuming. Technical correctness over social comfor…

  11

  12## The Response Pattern

  13

  14```

  15WHEN receiving code review feedback:

  16
• 04

  Refactoring

  Fait restructurer du code par l'agent sans changer son comportement : petits pas testés, un défaut à la fois, extraction, renommage, typage, patterns utiles.+

  sf-refactoring/SKILL.md, de github/awesome-copilot, licence MIT.

  1---

  2name: sf-refactoring

  3description: "Refactors code surgically to improve maintainability without changing behavior: small tested st…

  4---

  5

  6# Refactoring

  7

  8Improve the structure and readability of existing code without changing what it does. Refactoring is gradual…

  9

  10## When to use

  11

  12- Code that is hard to understand or change, functions or classes that are too large, smells that make the ne…

  13- "Clean up", "refactor", "simplify" requests.

  14

  15Not for: deepening module boundaries across a codebase (use `sf-architecture-deepening`), or reviewing someon…

  16
• 05

  Architecture Deepening

  Fait repérer par l'agent les modules superficiels à approfondir, les présenter dans un rapport HTML illustré, puis instruire le candidat choisi.+

  sf-architecture-deepening/SKILL.md, de mattpocock/skills, licence MIT.

  1---

  2name: sf-architecture-deepening

  3description: Scans a codebase (weighted toward recently changed hot spots) for deepening opportunities that t…

  4disable-model-invocation: true

  5---

  6

  7# Architecture Deepening

  8

  9Surface architectural friction and propose **deepening opportunities**: refactors that turn shallow modules i…

  10

  11This command is _informed_ by the project's domain model and built on a shared design vocabulary:

  12

  13- Read `references/design-vocabulary.md` for the architecture vocabulary (**module**, **interface**, **depth*…

  14- The domain language in `GLOSSARY.md` gives names to good seams; ADRs in `docs/adr/` record decisions this c…

  15

  16## Process
• 06

  Source-Grounded Code

  Fait fonder par l'agent chaque choix propre à un framework sur la doc officielle de la version installée, sources citées.+

  sf-source-grounded-code/SKILL.md, de addyosmani/agent-skills, licence MIT.

  1---

  2name: sf-source-grounded-code

  3description: Grounds every framework-specific implementation decision in official documentation for the exact…

  4---

  5

  6# Source-Grounded Code

  7

  8Every framework-specific code decision must be backed by official documentation. Don't implement from memory:…

  9

  10## When to Use

  11

  12- The user wants code that follows current best practices for a given framework

  13- Building boilerplate, starter code, or patterns that will be copied across a project

  14- The user explicitly asks for documented, verified, or "correct" implementation

  15- Implementing features where the framework's recommended approach matters (forms, routing, data fetching, st…

  16- Reviewing or improving code that uses framework-specific patterns
• 07

  TDD in Vertical Slices

  Fait coder l'agent test d'abord par tranches verticales : une frontière convenue, un test qui échoue, le minimum de code, sans tests couplés à l'implémentation.+

  sf-tdd-slices/SKILL.md, de mattpocock/skills, licence MIT.

  1---

  2name: sf-tdd-slices

  3description: Builds features and fixes bugs test-first in vertical slices (one agreed seam, one failing test,…

  4---

  5

  6# Test-Driven Development in Vertical Slices

  7

  8TDD is the red then green loop. This skill is the reference that makes that loop produce tests worth keeping:…

  9

  10When exploring the codebase, read `GLOSSARY.md` (if it exists) so test names and interface vocabulary match t…

  11

  12## What a good test is

  13

  14Tests verify behavior through public interfaces, not implementation details. Code can change entirely; tests…

  15

  16See [tests.md](tests.md) for examples and [mocking.md](mocking.md) for mocking guidelines.
• 08

  Diagnosing Bugs

  Fait diagnostiquer un bug difficile par l'agent : boucle de test qui échoue, repro minimale, hypothèses réfutables, test de non-régression.+

  sf-bug-diagnosis/SKILL.md, de mattpocock/skills, licence MIT.

  1---

  2name: sf-bug-diagnosis

  3description: "Runs a disciplined diagnosis loop for hard bugs and performance regressions: build a tight, det…

  4---

  5

  6# Diagnosing Bugs

  7

  8A discipline for hard bugs. Skip phases only when explicitly justified.

  9

  10When exploring the codebase, read `GLOSSARY.md` (if it exists) to get a clear mental model of the relevant mo…

  11

  12## Redact

  13

  14This skill has you show commands, outputs and captured artifacts. **Redact every secret first**: write `<REDA…

  15

  16If the redacted output is not enough to diagnose the bug, say so and ask the user.
• 09

  Root-Cause Debugging

  Impose à l'agent de trouver la cause racine avant toute correction : enquête, une hypothèse à la fois, test qui échoue, recul après 3 échecs.+

  sf-root-cause-debugging/SKILL.md, de obra/superpowers, licence MIT.

  1---

  2name: sf-root-cause-debugging

  3description: "Enforces root-cause investigation before any fix, through four phases: investigate (read errors…

  4---

  5

  6# Root-Cause Debugging

  7

  8**Core principle:** ALWAYS find root cause before attempting fixes. Symptom fixes are failure.

  9

  10**Violating the letter of this process is violating the spirit of debugging.**

  11

  12## The Iron Law

  13

  14```

  15NO FIXES WITHOUT ROOT CAUSE INVESTIGATION FIRST

  16```
• 10

  Security and Hardening

  Fait sécuriser le code par l'agent à partir du modèle de menace : entrées, injections, accès, sessions, secrets, dépendances, données perso.+

  sf-security-hardening/SKILL.md, de addyosmani/agent-skills, licence MIT.

  1---

  2name: sf-security-hardening

  3description: "Hardens code against vulnerabilities, threat model first: trust boundaries, STRIDE, abuse cases…

  4---

  5

  6# Security and Hardening

  7

  8Security-first development practices for web applications. Treat every external input as hostile, every secre…

  9

  10## When to Use

  11

  12- Building anything that accepts user input

  13- Implementing authentication or authorization

  14- Storing or transmitting sensitive data

  15- Integrating with external APIs or services

  16- Adding file uploads, webhooks, or callbacks

Extraits. Les dossiers complets sont livrés après le paiement.

Installation