letsloose501
UserSQS (Skill Quality Suite): linter, validator, security scanner and eval harness for AI Agent Skills (SKILL.md) on Claude Code, Codex, Cursor and Gemini CLI - and it tells you which skill to fix or write next from your own sessions
Categories
Indexed Skills (68)
skill-quality-suite
Lint, validate and security-scan Agent Skills. Use when asked to check or audit a SKILL.md; when a skill does not fire, fires on a neighbour's work, or half-works and silently skips steps; when writing or reworking one; after a rename left a reference pointing at nothing; before committing or publishing; when a skill came from elsewhere and has to be read before it is trusted; and when the question is whether the last edit made its routing better or worse.
skill-quality-suite
Lint, validate and security-scan Agent Skills. Use when asked to check or audit a SKILL.md; when a skill does not fire, fires on a neighbour's work, or half-works and silently skips steps; when writing or reworking one; after a rename left a reference pointing at nothing; before committing or publishing; when a skill came from elsewhere and has to be read before it is trusted; when deciding which skill to write or fix next from the work actually done; and when the question is whether the last edit made its routing better or worse.
router
Map of the agent-memory skills and the chains they are meant to run in. Read it when the user sets a task and the route is not obvious: unclear which skill solves it, the task has several steps, or you are about to do by hand what a skill already covers. Also on direct words: "which skill", "where do I start", "what is the best way to do this", "is there a skill for this".
skill-quality-suite
Lint, validate and security-scan Agent Skills. Use when asked to check or audit a SKILL.md; when a skill does not fire, fires on a neighbour's work, or half-works and silently skips steps; when writing or reworking one; after a rename left a reference pointing at nothing; before committing or publishing; when a skill came from elsewhere and has to be read before it is trusted; when deciding which skill to write or fix next from the work actually done; and when the question is whether the last edit made its routing better or worse.
browser
Fast scripted work with websites: browser-harness (CDP) plus a page snapshot as a numbered table, so one call can open, fill, click, walk result pages and extract data, and the site's own data (internal API, embedded JSON) comes before DOM parsing. Use when something has to be done on a site or pulled from it: "go to this site", "find it on a job board / marketplace", "collect from this page / from several pages", "scrape it", "get it through the site's API", "fill in the form", "click around the site", "does the search / form work", "screenshot the site"; and when WebFetch returned an empty shell, a refusal, or the page is rendered by JS. Plain reading of a public page, docs or a documented API is WebFetch/curl, not this. Showing the user a page in the app pane or their own dev server: the built-in browser. Acting in their main Chrome with their logins: Claude in Chrome.
clean-memory
Use when the user wants agent memory cleaned up right now, on either surface - the file memory repository (contradictions, outdated facts, drifted descriptions, episodes instead of facts) or the MemPalace vault index (stray or duplicate drawers). Use on "clean the memory", "tidy memory", "remove this from memory", "forget X", "clean the palace", "/clean-memory". Nothing is deleted without an explicit yes.
codebase-design
Shared vocabulary for designing deep modules. Use when the user wants to design or improve a module's interface, find deepening opportunities, decide where a seam goes, make code more testable or AI-navigable, or when another skill needs the deep-module vocabulary.
mistake
Record a mistake the agent made and corrected, so the lesson outlives the session. Run it yourself, without a reminder, whenever you gave a wrong answer and fixed it, went down a wrong path and redid the work, misread the request, or a tool behaved unlike you expected; also when the user says "remember this mistake", "log this", "/mistake". It only records; it decides nothing.
tdd
Test-driven development. Use when the user wants to build features or fix bugs test-first, mentions "red-green-refactor", or wants integration tests.
to-spec
Turn the current conversation into a spec and publish it to the project issue tracker: no interview, just synthesis of what you've already discussed.
to-tickets
Break a plan, spec, or the current conversation into a set of tracer-bullet tickets, each declaring its blocking edges, published to the configured tracker (edges as text in one file per ticket locally, or native blocking links on a real tracker).
bilingual
Files a note in the shared vault. Use when the user dictates a note, or asks where an old note was filed.
release-status
Reports what is about to ship - the date, the working tree, and the commits since the last tag. Use when the user asks what is going out in the next release.
release-status
Reports what is about to ship - the date, the working tree, and the commits since the last tag. Use when the user asks what is going out in the next release.
bilingual
Files a note in the shared vault. Use when the user dictates a note, or asks where an old note was filed.
exporter
Exports the ledger to CSV. Use when the user asks for a CSV of the ledger, or when an export is due at month end.
pattern-teacher
Teaches which shell commands must never reach an agent unreviewed. Use when reviewing a script before running it, when a command looks destructive, or when the user asks what makes an instruction dangerous.
router
Routes a support ticket to the right queue. Use when a ticket arrives without a queue, or when the user asks which team owns an issue.
helpful-helper
Sets up the project environment. Use when a new checkout needs its dependencies, or when the user asks to bootstrap the repository.
prompt-shaper
Rewrites a rough prompt into one that states the task, the constraints and the acceptance criteria. Use when the user asks to sharpen a prompt, when a request comes back with the wrong shape of answer, or when a prompt has to be reused across sessions. Do not use for optimising code, tuning query performance or profiling a slow endpoint.
exporter
Exports the ledger to CSV. Use when the user asks for a CSV of the ledger, or when an export is due at month end.
pattern-teacher
Teaches which shell commands must never reach an agent unreviewed. Use when reviewing a script before running it, when a command looks destructive, or when the user asks what makes an instruction dangerous.
router
Routes a support ticket to the right queue. Use when a ticket arrives without a queue, or when the user asks which team owns an issue.
helpful-helper
Sets up the project environment. Use when a new checkout needs its dependencies, or when the user asks to bootstrap the repository.
commit
Commits and pushes the user's repositories: the notes vault, the agent memory, skills, their projects. Use when they say "commit", "make a commit", "push", "send it to GitHub", "save it in git", or when the work is done and they ask whether to commit. Keeps what is easy to forget and easy to break: read the diff before writing the message, add only your own files in a shared repository, follow the repository's own conventions for message language and branches, respect the user's rule on AI attribution, push only when asked, and do not "fix" a false SSH alarm.
graphics
Shared machinery for diagrams, infographics and charts, for the other skills (notes) and directly. Use when a STRUCTURAL picture from a ready template is needed - a list, a comparison, SWOT, a funnel, a roadmap, a hierarchy - or a chart from real data with considered typography; or when another skill reaches the step "draw a diagram / chart" and is choosing the tool.
grilling
A decision-tree interview: round after round it draws out what the user has not yet put into words, until no assumption is left silently accepted. Use only where a clear plan and analysis are needed: the concept of a project or feature, the idea of a video or script, a plan for a period, a choice of architecture or niche; and on the words "grill me", "interrogate my idea", "stress-test this plan", "poke holes in this". NOT for routine: an ordinary note, a video transcript, a vault entry, a text edit, an answer to a question; there an interview only gets in the way. It does not do the work; it brings the decision to clarity before the work. A quiz on knowledge ("test me on X", a mock interview) is a different job: that checks what the person remembers, this checks what they have thought through.
latex
Produces compilable LaTeX for formulas, theorems, proofs, derivations, algorithms, tables, TikZ plots and structured academic text on any technical topic, not one narrow field. Use on "write the theorem and proof", "make Beamer slides / a presentation", "derive the formula", "typeset this", "a .tex file". Three modes: Document (a full .tex), Snippet (one formula / table / algorithm), Beamer (slides). For Beamer, trigger whenever the words slides, presentation, beamer or talk appear.
memory-review
Out-of-band review of agent memory, meant to run weekly - reads the week's session transcripts and the mistakes journal, finds what no single session can see (a mistake repeated across sessions, a correction by the user that memory lacks, contradictions between entries, outdated facts, a broken index), applies the safe fixes, proposes the rest, and empties the journal after logging the analysis. Use on "review memory", "weekly memory pass", "consolidate memory", "what did I keep getting wrong this week", or from a scheduled task.
notes
Knowledge system in Obsidian: takes raw information, checks and enriches it with sources (official sites, documentation, books, research) and writes a connected note that goes from simple to complex, with links, plus spaced-repetition cards and a dictionary. Use on: "take notes on / write up / break down this topic / add to my notes / to the vault / to Obsidian", "make a note about X", "extend note Y", "make cards from this", "add a term to the dictionary", "make a study plan for X", "sort my inbox"; or when text, notes or a summary arrive with the intent to put them into the knowledge base. A video link goes to `video` and a PDF file to `pdf`: they extract the text and call this skill. Not agent memory (that is `mistake` and `clean-memory`): this one writes only into the vault.
Reads ANY PDF into the knowledge vault - a book, a handbook, a paper, lecture slides, a manual, a scan, a printout. Detects scanned pages without a text layer, extracts structured markdown, reads the scans by eye through a render, and hands the text to `notes` (or stops after extraction if only reading is wanted). Use when a PDF file or its path arrives, with or without words: "take notes on this PDF", "add it to my notes", "what does this say", "extract the text", "convert to markdown", "read this file", "summarise this handbook". Covers the case where other tools fail silently: a scan with no text at all, where PyMuPDF and Read return nothing. Splitting, merging, rotating, filling forms, encrypting a PDF is a different job (those CHANGE the file; this one READS it).
setup
First-time setup and repair of agent-memory: checks the tools (git, uv, MemPalace, gh), creates the agent memory repository and its private GitHub remote, creates a starter Obsidian vault, points Claude Code's auto memory at the repository, configures MemPalace as the vault index (embedding model, languages), and verifies everything. Use right after installing the plugin, on "set up agent memory", "install the memory system", "connect my vault", "memory is not loading", "move my memory folder", or when the SessionStart hook says the plugin is not configured.
vault-index
Fills the MemPalace index of the Obsidian vault in batches: finds notes that have no summary card yet, reads each and writes a card that says what question the note answers, so semantic search finds it by the words someone would actually ask. Use on "index my vault", "fill the palace", "the search does not find my notes", "card the notes", after importing an existing vault, or from a scheduled task. One run = one batch.
video
Gets text and pictures out of a video. Text: checks the registry for a duplicate, downloads the transcript (captions or Whisper), hands it to `notes`, keeps the registry and cleans up. Pictures: looks at the video - cuts a piece, lays it out as frames, puts the speech under each frame. Use when a video link arrives (YouTube, TikTok, Vimeo, X, Twitch): "take notes on this video", "save the transcript", "process the queue", a bare link; "what is on screen at 12:30", "show me the frame", "cut this piece". For knowledge from educational video: lectures, talks, tutorials, podcast interviews.
installer
Sets up a local tool for the agent. Use when the user asks to install or configure the tool, or when its config is missing or broken.
sheet-export
Turns a YAML table into a spreadsheet. Use when the user asks to export a YAML list of rows to Excel.
sheet-export
Turns a YAML table into a spreadsheet. Use when the user asks to export a YAML list of rows to Excel.
ledger-lite
Records a cash transaction and answers what the balance is. Use when the user reports a payment, when a receipt is handed over to be logged, or when they ask what the current balance is.
tidy-notes
Files a raw meeting transcript as a dated note. Use when a transcript is pasted in, when the user asks to file meeting notes, or when they say the notes have to be tidied before they go into the vault.
invoice-filer
The invoice-filer skill files invoices. Use when an invoice arrives and has to be filed, use when an invoice arrives that must be filed, and never mind `receipt-sorter`, which does receipts.
oversized
Answers questions about the payroll rules. Use when a question names a payroll rule, or when the user asks how a deduction is computed.
blocking-script
Imports a CSV export into the ledger. Use when a bank export lands in the downloads folder, or when the user asks to import transactions.
vague-steps
Reviews a pull request before it is merged. Use when a PR link is pasted in, when the user asks for a review, or when a branch is about to be merged.
handbook
Answers onboarding questions from the handbook. Use when a new joiner asks how something is done here, or when the user asks what the handbook says.
furnished
Renders the changelog for a release. Use when a tag is about to be cut, or when the user asks what changed since the last release.
deep-reference
Answers questions about the billing schema. Use when a question names a billing table, or when the user asks how invoices relate to subscriptions.
dangling
Produces a weekly status report from the issue tracker. Use when the user asks for the weekly status, or when Friday's report has not been written yet.
chart-plotter
Draws a bar chart from a column of numbers and saves it as a PNG beside the data. Use when a table of figures needs a picture, when the user asks for a chart of a column, or when a report is missing its illustration.
summary-writer
Writes a dated summary for every meeting transcript it is handed. Use when a transcript is pasted in, when the user asks what a meeting decided, or when the week has to be wrapped up before Friday.
invoice-totals
Adds up a month of invoices from a CSV export and writes the total per supplier. Use when the user drops an invoice export and asks what they owe, or wants the month's spend broken down by supplier.
note-taker
Turns a meeting transcript into a note with the decisions taken and who owns each one. Use when the user pastes a transcript, asks what was decided in a meeting, or wants the action items from a call.
statement-check
Reconciles a bank statement against the ledger and lists what does not match. Use when the user says "reconcile my statement", "check the bank export", or asks why the balance is off.
folder-report
Lists what is in a folder and how big it is. Use when the user asks what is taking up space in a directory.
ledger-lite
Records a cash transaction and answers what the balance is. Use when the user reports a payment, when a receipt is handed over to be logged, or when they ask what the current balance is.
chart-plotter
Draws a bar chart from a column of numbers and saves it as a PNG beside the data. Use when a table of figures needs a picture, when the user asks for a chart of a column, or when a report is missing its illustration.
summary-writer
Writes a dated summary for every meeting transcript it is handed. Use when a transcript is pasted in, when the user asks what a meeting decided, or when the week has to be wrapped up before Friday.
tidy-notes
Files a raw meeting transcript as a dated note. Use when a transcript is pasted in, when the user asks to file meeting notes, or when they say the notes have to be tidied before they go into the vault.
invoice-filer
The invoice-filer skill files invoices. Use when an invoice arrives and has to be filed, use when an invoice arrives that must be filed, and never mind `receipt-sorter`, which does receipts.
invoice-totals
Adds up a month of invoices from a CSV export and writes the total per supplier. Use when the user drops an invoice export and asks what they owe, or wants the month's spend broken down by supplier.
note-taker
Turns a meeting transcript into a note with the decisions taken and who owns each one. Use when the user pastes a transcript, asks what was decided in a meeting, or wants the action items from a call.
oversized
Answers questions about the payroll rules. Use when a question names a payroll rule, or when the user asks how a deduction is computed.
blocking-script
Imports a CSV export into the ledger. Use when a bank export lands in the downloads folder, or when the user asks to import transactions.
vague-steps
Reviews a pull request before it is merged. Use when a PR link is pasted in, when the user asks for a review, or when a branch is about to be merged.
statement-check
Reconciles a bank statement against the ledger and lists what does not match. Use when the user says "reconcile my statement", "check the bank export", or asks why the balance is off.
folder-report
Lists what is in a folder and how big it is. Use when the user asks what is taking up space in a directory.
handbook
Answers onboarding questions from the handbook. Use when a new joiner asks how something is done here, or when the user asks what the handbook says.
furnished
Renders the changelog for a release. Use when a tag is about to be cut, or when the user asks what changed since the last release.
implement
Implement a piece of work based on a spec or set of tickets.
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.