bastien-gallay

Analyse rhétorique et épistémologique d'articles, discours et textes argumentatifs. Utiliser ce skill quand l'utilisateur demande d'analyser la qualité argumentative d'un texte, d'identifier des sophismes ou biais, d'évaluer la fiabilité des sources citées, de déconstruire la logique d'un raisonnement, ou de produire une réécriture critique structurée d'un document.

Analyse rhétorique et épistémologique d'articles, discours et textes argumentatifs. Utiliser ce skill quand l'utilisateur demande d'analyser la qualité argumentative d'un texte, d'identifier des sophismes ou biais, d'évaluer la fiabilité des sources citées, de déconstruire la logique d'un raisonnement, ou de produire une réécriture critique structurée d'un document.

Mise à jour totale ou partielle de la revue multi-axes du site (éditorial, visuel, tech, prose — plus perf en demi-axe et mesure/pilotage prototypé inactif), avec synthèse transverse vivante, vérification adversariale, reprise incrémentale et relecture des passes précédentes. Trigger : l'utilisateur tape `/revue`, demande à « rejouer la revue », « mettre à jour la synthèse », « relancer l'axe visuel/tech/éditorial/prose », ou « où en est la revue ? ».

Close out a session that reached a milestone — a feature landed, a bug fixed, an investigation concluded, a document consolidated — by running a 5S pass over the work: Seiri (sort what stays, what goes), Seiton (put each thing in its place), Seiso (clean — reconcile every document carrying a now-superseded claim, run the checks), Seiketsu (standardise — promote the durable lesson up to where the next reader will meet it), Shitsuke (sustain — leave an entry point for whoever arrives next). Reads project config from `.wrap.md` at project root or `.personal/wrap/config.md`; writes one if absent. Not a daily ritual — it triggers on a milestone, not on a clock. Trigger: the user types `/wrap`, says "wrap up", "close out this milestone", "on a fini, range", "finalise this session", or signals that a piece of work has reached a stopping point worth leaving clean.

Strategic advisor for market analysis, positioning, traction, and project communication. Use when the user wants to analyse a market opportunity or a competitive landscape, find the right communication angle for a project (commercial or not), build a value proposition or a pitch, identify traction channels, work on personal branding or thought leadership, or apply a strategic framework (Blue Ocean, PESTEL, BMC, SONCAS, 19 traction channels, Wardley Maps, Jobs-to-be-Done, Obviously Awesome, and more). Also trigger on "marché", "positionnement", "traction", "pitch", "concurrents", "communication projet", or when the user asks how to reach an audience, stand out, find clients, or grow visibility — including for non-commercial projects (open source, community, indie games, professional expertise). Trigger when the user asks for an original or unusual framework angle.

Read a faceto event-storming board back to a stuck modeller and propose the missing pieces as reviewable events. Use when working in a repo that holds a `*.event-log.jsonl` (a faceto board) and the user asks to narrate the board, tell its story, find what's missing / what happens next, spot gaps or suspicious shapes, or "get me unstuck" mid session. Reads the log directly and — only on explicit approval, one at a time — appends proposals through a running `faceto serve` via `POST /comment` (server mints ids). Not for editing faceto's Rust, the model.json source, or the log file directly.

<What the skill does, in one or two sentences.> Use when <the concrete trigger phrases a user would type or the situation that should invoke this skill> — the trigger clause is what makes the skill discoverable, so make it specific.

Builds a CLI Python tool to automatically extract relevant screenshots from a meeting video, using a transcript and user-defined topics. Combines ffmpeg for frame extraction with Claude vision for semantic validation in an iterative feedback loop. Use this skill whenever the user wants to extract screenshots or frames from a video based on topics, themes, or transcript passages — even if they don't say "skill" or "CLI". Trigger on: "extract screenshots from video", "capture frames from meeting", "find slide in recording", "screenshot at timestamp", "video + transcript + topics", "grab frame when X is discussed". Also trigger when the user provides a video file and a transcript and wants to find visual moments matching descriptions.

Work one issue (or a scoped task) as an adversarial pair, not a delegation — the goal is to reduce the TRIPLE debt (technical, cognitive, intent), keeping the human the author of their own system. The human produces positions; the AI attacks them; the human revises. Test-first with a per-test Red→Green→Refactor→Reflect loop, two pauses, async gates that interrupt where it has leverage, a post-process retention pass (visual recap + novelty ledger), and emit-now/ store-later measurement hooks. Use when the user types `/pair-with-me #<n>`, asks to "pair on", "work this issue with me", "co-create / co-engage on a fix", or wants help on an issue WITHOUT being turned into a reviewer of AI output they can't defend.

Facilitate a structured brainstorming session with adaptive technique sequencing — MoSCoW, Impact/Effort, Six Hats, SCAMPER, and more from a hot-swappable technique library. Use when the user types /brainstorm, or asks to "brainstorm", "facilitate ideation", "run a structured brainstorm", "help me think through options", or wants adaptive technique sequencing for idea generation and decision-making. Trigger: brainstorm, ideation, facilitate a session, structured brainstorm, generate ideas.

Critical, sourced counter-assessment ("contre-expertise") of a working method described by the user (articles, notes, practice README): extraction of the model into falsifiable propositions, author interview, multi-agent deep research (comparable methods, practices, cited sources), two-stage adversarial verification, rival method (challenger steelman), report with dated perishable verdicts, and a capped action plan in experimental format. Progresses incrementally by default (resumes from on-disk artifacts). Trigger: the user types /contre-expertise, asks to "audit my method", "critically review the way I work", "stress-test my practices against the state of the art", « passer ma méthode au crible », « analyse critique de ma façon de travailler », or wants their practice confronted with the state of the art without flattery.

Boucle d'édition critique pour écrits longs (articles, billets, essais). Le facilitateur (Claude) ne rédige PAS les paragraphes de prise de position — il critique, structure, propose des squelettes, et débusque les pièges récurrents (thèse molle, hedging, registre LinkedIn, re-cloisonnement, énumération qui esquive la nomination). Travail paragraphe par paragraphe : l'utilisateur écrit, Claude critique, l'utilisateur réécrit, Claude reporte dans le fichier. Trigger : l'utilisateur tape `/redaction`, demande à « écrire un article », « relire ce que je viens d'écrire », « critiquer ce paragraphe », ou veut aider à finaliser un brouillon long.

Job offer analysis, profile/role fit assessment, and tailored CV/cover letter generation. Activate this skill when the user mentions a job application, a job posting, a cover letter, or an adapted CV.