Claude Skills by adamlinscott

Deterministic close-out gate for a session or work tree. Sweeps a fixed set of sources — the todo list, this session''s own promises, git state, TODO/FIXME markers the diff introduced, test and build state, and any issue or ticket referenced — and checks that every open point has been explicitly decided: done, won''t do, deferred, or handed to another session. Strict by default, with a severity floor that lifts only on evidence of how the project actually works (CI on every push, a user who c...

Surface and confirm the load-bearing assumptions behind a task before a long or expensive run — goal, root, what may be edited, what is off-limits, what done means, and open questions — separating what can be cited from what is being guessed, so bad direction is caught before time is burned. Also pressure-tests the plan's acceptance criteria, flagging any that are weak or built on unproven assumptions, and asks the blocking uncertainties (scope boundaries, interpretations, soft "done" bars) a...

Print a plain-English re-entry briefing for a user who has lost the thread of a long-running session — after a distraction, overnight, or over a weekend. Re-explains what the session is, where it got to, what state the work is in right now, what Claude is waiting on from the user, and the single next step. Grounds every claim in git and files on disk before narrating, so it re-explains the session rather than inventing a tidy story of it. Report only: it writes nothing and never resumes work....

Check whether the installed copy of this skills collection is behind its remote, and offer to update it. Runs a script that resolves the clone the skills are linked from, respects a once-a-day rate limit and a remembered refusal, and reports one of: up to date, N commits behind, unsafe to touch, or offline. Only "behind" is ever mentioned to the user; everything else is silent. On a yes it pulls and re-links, changing nothing about which skills are installed. Use when the user asks whether th...

Audits a repository's Claude context-injection setup — CLAUDE.md, CONTEXT.md, docs/, .claude/agents/, and the per-project memory directory. Reports bloat, broken links, orphaned docs, security risks, missing rules in subagent prompts, and conflicts between memory and project instructions. Use when the user asks to audit their Claude setup, asks "what's wrong with my CLAUDE.md", wants to know if their agents/memory/docs are configured well, or wants to improve Claude's effectiveness in this re...

BETA / under development. Mine the current project's Claude Code sessions for recurring corrections, then interrogate them — consolidate clusters into themes, ask a SET of open "why" questions per theme, self-answer the evidence-answerable ones and forward the developer-only ones to the user, record confirmed answers, and optionally export a CLAUDE.md of principles. Use when the user says "/debrief", "debrief my sessions", "what patterns do you see in how I work", or wants to refresh / interr...

Fresh-context audit of a finished diff against the conversation's stated intent. A subagent with zero conversation history reviews the diff blind — reporting what it believes the change does, its quality, and any oversights — then the main context reconciles that blind read against what the work was actually meant to do, and surfaces any assumption about real-world behaviour that the code cannot settle. Use when a chunk of work is complete and you want fresh eyes to confirm nothing was missed...

Run a settled implementation goal as a bounded autonomous build loop — lock the goal from the conversation and any docs written this session, front-load every decision, map the terrain, write a completion-invariant contract before any code, then loop (building with explicit Agent-tool fan-out, committing at intervals, verifying at milestones with fresh-eyes against the contract) until the invariants hold, and close out. Gated on a `--confirm` flag asserting the user has manually set ultracode...

Audit Claude's per-user memory for this project and produce a report. Asks at the start whether the user wants a non-technical summary or a full technical audit. Never edits memory — produces a report only. Usage: /memory-audit

Turns "this is broken" into a well-formed issue in the team's tracker, written in the codebase's own vocabulary. Built for someone who uses the product and does not read code — it asks a few plain-English questions, works out for itself which part of the code they are describing, checks whether the same thing has already been reported, and files it, but only after they say yes. Use when the user says something is broken, looks wrong, isn't working, is behaving oddly, seems like a bug, or asks...

Halt a drifting agent and re-anchor it to codebase evidence for the current task, clearing speculative context without a full compaction. Use when the agent has gone off course, hallucinated files/APIs/behavior, or started over-building beyond what was asked — or when the user types /reground or says "stop", "re-ground", "re-crown", "I don't see evidence for that", or "you're building something we don't need".

Alias for the raise-issue skill. Turns "this is broken" into a well-formed issue in the team's tracker, written in the codebase's own vocabulary — a few plain-English questions, a check of which part of the code is being described, a duplicate check, then filing it once the person says yes. Built for someone who uses the product and does not read code. Invoke with /report-issue.

Sets up the security foundation that lets you run Claude Code fast — auto mode, custom tools, minimal interruptions — without it doing something irreversible you never asked for. Two modes. Developer, a permissive default plus a thin deny-only brake on the genuinely irreversible. Vibe, deny-by-default for people who don't know git, so code changes flow freely while merges, force-pushes, deploys, secret reads and spend are blocked at the Claude layer. Grounds itself in the repo's existing Clau...

Alias for the seatbelt skill. Sets up the security foundation that lets you run Claude Code fast — auto mode, custom tools, minimal interruptions — without it doing something irreversible you never asked for. Two modes, one for developers and one for non-technical vibe coders. Use when setting up a project for AI-assisted development, when handing a repo to a non-technical builder, or when the user says they want Claude to stop breaking things, worries it will delete their work, or wants to r...

[skips confirmation] The unattended form of /ship-it. Gets this conversation''s work merged and deployed to every environment the repo has, production included — and does NOT stop to confirm when the release also carries unreleased work by other developers or other agent sessions. Identical to /ship-it in every other respect: same grounding in branches, PRs and environments, same verification, same merge-conflict policy. Takes no arguments by default; an optional scope override such as "only ...

Get the work done in this conversation all the way out — merged, and deployed to every environment the repo has, production included. Grounds itself first in the branches written during this session, the open and closed PRs behind them, and the environments and CI/CD workflows the repo actually defines, then lands and releases the work and verifies each environment is running the commit. If the release would carry other people''s unreleased changes along with it, summarises whose work rides a...

Alias for the ttp skill. Shape user-facing output to be brief and direct — lead with the substance, keep the default answer short, and expand only when the user asks for detail. Leaves the user in control: Claude settles small, reversible, or already-decided points and proceeds, but routes high-impact calls (ADR-worthy, production, project shape, which features get built, infrastructure) to the user with enough context to decide. Only shapes prose written for the user to read; Claude''s own r...

Shape user-facing output to be brief and direct — lead with the substance, keep the default answer short, and expand only when the user asks for detail. Leaves the user in control: Claude settles small, reversible, or already-decided points and proceeds, but routes high-impact calls (ADR-worthy, production, project shape, which features get built, infrastructure) to the user with enough context to decide. Only shapes prose written for the user to read; Claude''s own reasoning, tool use, code,...