个人作者
Sebastien Rousseau
github.com/sebastienrousseau ↗
Payments leader, 19 years in banking. I build open-source ISO 20022, post-quantum cryptography and applied AI for finance.
- Skills
- 39
- 仓库
- 2
- 职业覆盖
- 0
- 场景覆盖
- 0
- Stars
- ★ 1
Sebastien Rousseau 的 Skills 39
example-skill
Use when the user needs this reusable workflow. Replace this with trigger-rich wording that says when to load the skill, what task it handles, and when not to use it.
writing-plans
Turn a vague or multi-step request into an ordered list of small steps, each with its own done-condition and verification. Load before starting work that touches several files, has unclear scope, or spans more than one session, and when a request needs decomposing before any code is written. Produces a plan a reviewer can check against, not a to-do list.
web-research-and-source-triage
Use when gathering current web evidence, comparing sources, triaging credibility, reading GitHub repositories, checking docs, or deciding whether a claim needs live verification.
vetting-a-skill-before-install
Use before installing, importing or enabling a third-party skill, plugin, marketplace or skills repository in Claude Code, Codex, Antigravity, Cursor or any agent. Pins the source to a commit, audits every skill and the agent configuration it ships, reads what the audit could not cover, and ends in install, mirror or refuse. Not for auditing this registry's own skills.
verification-before-completion
The gate between doing work and claiming it is done. Load before saying finished, fixed, working, or ready, before marking a task complete, and before reporting an outcome to a human. Converts each claim into the observation that would justify it, and names the rationalizations that let unverified work ship. Not the project's own evidence bar or test taxonomy.
using-agtmls
Meta-router for the AgtMLS skill catalog. Read first when unsure which AgtMLS skill applies, when several look relevant, when asking 'is there a skill for porting or noyalib work', or when onboarding to how the hub's skills and shared references fit together.
test-driven-development
Write the failing test first, then the code that makes it pass. Load when adding behaviour, changing behaviour, or pinning a bug, and whenever a test was authored after the code it covers. Enforces red-green-refactor with an observed red phase, so a test that never failed is never mistaken for proof.
systematic-debugging
Find the root cause of a defect before changing anything. Load when something crashes, hangs, returns the wrong value, fails intermittently, or broke after a change, and the cause is not yet known. Drives reproduce, localize by bisection, explain the mechanism, then fix with a failing-first regression test. Replaces guess-and-check editing.
refactoring-safely
Change the shape of code without changing what it does. Load when you extract, rename, move, deduplicate, or restructure a module safely across its call sites, and whenever a diff mixes restructuring with new behaviour. Requires a characterisation net before the first edit and keeps structural and behavioural changes in separate commits.
receiving-code-review
Respond to review feedback on your own change without reflexively complying or reflexively defending. Load when handling a change request on your own PR, when a reviewer leaves comments, disagrees with an approach, or asks for something out of scope, and when deciding whether to argue a point or accept it. Every comment gets a decision and a reply; silent edits and silent skips are the failure modes.
pr-review-and-release
Use when preparing a pull request, judging release readiness, reviewing implementation risk with release discipline, writing release notes, checking CI, packaging artifacts, or deciding whether a repository is ready to ship.
incident-response
Restore service first, find the cause afterwards. Load when something is broken in production right now: an outage, a bad deploy, data at risk, alerts firing, users affected. Prioritises mitigation over diagnosis, keeps a running timeline, and defers the root cause to a blameless follow-up. Not for unhurried debugging of a reproducible defect.
handoff
Write the note that lets someone else — or a later session — resume work in progress without re-deriving it. Load when context is running out, when pausing unfinished work, when switching who or what continues it, and when asked where things stand. Records tree state, evidence, the next concrete action, and the dead ends. Not a summary of what you did.
giving-code-review
Review someone else's change so the feedback is actionable and proportionate. Load when reading a pull request, diff, or patch you did not write, and when deciding whether a concern is blocking or merely a preference. Prioritises correctness over style, labels every comment with its severity, and separates what must change from what you would have done differently.
cross-language-port
Port a module, CLI, or function between Rust, Python, Go, C++, Swift, TypeScript, JavaScript, Ruby, and Bash - preserving logic, adopting target idioms, and proving equivalence with a differential golden-I/O harness. Use for 'port/rewrite this in <lang>'. Not micro-snippets or redesigns.
brainstorming
Turn a vague request into a specification before any plan or code exists. Load when an ask is underspecified, when a feature request arrives with no acceptance criteria, or when you notice you are guessing at what someone wants. Surfaces the decisions, constraints, and deliberate non-goals by asking rather than assuming. Not for diagnosing a defect.
authoring-portable-skills
Use when writing or editing a SKILL.md that more than one agent will read (Claude Code, Codex, Antigravity, Cursor, Copilot, Gemini CLI, OpenCode). Keeps the shared file to what every agent honours, sizes the body, checks it with validate-skills.py and the agents themselves, and requires evidence before calling it hardened. Not for choosing what a skill should teach.
anti-slop-pr-and-writing
Eliminate AI slop, clichés, sycophancy, and robotic filler from PR descriptions, commit messages, code comments, and technical documentation. Load when reviewing or authoring PRs, drafting release notes, or stripping conversational apologies and binary contrasts while preserving human voice and technical facts.