Back to search
laravel-logic-trace-tests
SkillGitHub
Agent Skill: Laravel tests from evidence. Traces request flow, probes edge cases, records CURRENT vs RECOMMENDED behavior
Safety scan
A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.
Badge for your README
Markdown
[](/listings/suliman-al-helou--laravel-logic-trace-tests)HTML
<a href="/listings/suliman-al-helou--laravel-logic-trace-tests"><img src="/api/listings/suliman-al-helou--laravel-logic-trace-tests/badge.svg" alt="findsafeskills safety rating" /></a>Install
i
Install instructions from the README
As written by the repo's author. findsafeskills has not verified these steps.
Replace ` Suliman-Al-Helou` with the account that hosts this repo.
**Gemini CLI**
gemini skills install https://github.com/ Suliman-Al-Helou/laravel-logic-trace-tests.git \ --path skills/laravel-logic-trace-tests --scope workspace --consent
**Any agent that supports `npx skills`**
npx skills add Suliman-Al-Helou/laravel-logic-trace-tests --skill laravel-logic-trace-tests
**Claude Code (manual)**
git clone --depth 1 https://github.com/ Suliman-Al-Helou/laravel-logic-trace-tests.git /tmp/lltt mkdir -p .claude/skills && cp -r /tmp/lltt/skills/laravel-logic-trace-tests .claude/skills/
**Laravel Boost**
mkdir -p .ai/skills && cp -r /tmp/lltt/skills/laravel-logic-trace-tests .ai/skills/ php artisan boost:update
Copy the whole folder: the scripts and references live beside `SKILL.md`.
What was scanned
The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.
README.md · README · 105 lines
- # laravel-logic-trace-tests
- An [Agent Skill](https://github.com/anthropics/skills) for Laravel/PHP back-end tests that are built
- from **evidence**, not guesses.
- It makes the agent trace the real request flow (route → middleware → FormRequest → controller →
- service → model → events/jobs → policy), find edge cases by reading the code and running probes,
- and then write tests that pin the **current** behavior and state the **recommended** behavior with
- a one-line reason.
- ## What you get
- For each edge case, a comment block above the test (checked by a gate script):
- ```php
- // EDGE: E-03
- // COVERS: D-07
- // CURRENT: second identical POST returns 201 and creates a duplicate order
- // RECOMMENDED: 409 and no duplicate (precedent: app/Http/Controllers/PaymentController.php:58)
- // WHY: client retries after a timeout would double charge
- // EVIDENCE: app/Services/OrderService.php:42 ; probe P-03
- ```
- Plus `.trace/logic-map.md`, `.trace/decision-table.md`, `.trace/probes/*.json` and
- `.trace/behavior-report.md`.
- ## Rules the skill enforces
- 1. No edge case without a `file.php:line` the agent opened and a probe result it actually ran.
- 2. `RECOMMENDED` comes from a project precedent or Laravel docs, otherwise `none established`.
- 3. Application code is never modified. Only `tests/` and `.trace/`.
- 4. Probes run through the test runner only (test DB), never against the default environment.
- 5. A gate script fails the run on incomplete edge blocks, missing evidence, assertion-less tests,
- and uncovered decision-table rows.
- Designed to stay usable with small/cheap models: short `SKILL.md`, one phase per reference file,
- mechanical checks instead of prose rules.
- ## Install
- Replace ` Suliman-Al-Helou` with the account that hosts this repo.
- **Gemini CLI**
- ```bash
- gemini skills install https://github.com/ Suliman-Al-Helou/laravel-logic-trace-tests.git \
- --path skills/laravel-logic-trace-tests --scope workspace --consent
- ```
- **Any agent that supports `npx skills`**
- ```bash
- npx skills add Suliman-Al-Helou/laravel-logic-trace-tests --skill laravel-logic-trace-tests
- ```
- **Claude Code (manual)**
- ```bash
- git clone --depth 1 https://github.com/ Suliman-Al-Helou/laravel-logic-trace-tests.git /tmp/lltt
- mkdir -p .claude/skills && cp -r /tmp/lltt/skills/laravel-logic-trace-tests .claude/skills/
- ```
- **Laravel Boost**
- ```bash
- mkdir -p .ai/skills && cp -r /tmp/lltt/skills/laravel-logic-trace-tests .ai/skills/
- php artisan boost:update
- ```
- Copy the whole folder: the scripts and references live beside `SKILL.md`.
- ## Use
- Run from the Laravel project root and name ONE feature or endpoint:
- > Write tests for `POST /api/orders`. Use the laravel-logic-trace-tests skill.
- ## Requirements
- - Laravel project with Pest or PHPUnit and a configured test database
- - `bash`, `php`, `python3` on PATH
- ## What the scripts do (read before installing)
- - `scripts/detect_env.sh` — read-only: runs `php artisan --version`, greps `phpunit.xml` / `composer.json`,
- counts test files, writes `.trace/env.txt`.
- - `scripts/verify_tests.py` — static checks over `tests/` and `.trace/`. With `--run` it executes
- `php artisan test`.
- - The skill declares no `allowed-tools`.
- ## Known limitations (v0.1.0)
- - Not yet evaluated on multiple models; compliance of cheap models is unmeasured.
- - The gate is regex-based and can miss unusual test layouts (data providers, dynamic tests).
- - The gate does not yet reject a `RECOMMENDED` precedent that points at the same line as `EVIDENCE`
- (circular evidence).
- - The gate does not detect gaps in edge IDs (an edge dropped without being reported as UNVERIFIED).
- - Speed: probes and full-suite runs can make a run take several minutes.
- ## Roadmap
- - Circular-evidence check in the gate
- - Edge-ID continuity check
- - Batched probes and focused test runs
- - Evals with with-skill vs without-skill runs on a small model
- ## License
- MIT
Ratings and comments
Sign in to rate this skill.
Sign inFree to use. If it’s helped, pass it on.