Loading…
Not enough history yet: 3 of 7 days recorded. Until then this list is ordered by the most recent content change.
77205 skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| leak-triageDavidCho1999/archaiflow-skillsWorks out what kind of water is appearing before anyone opens a wall: rain driven in, groundwater, a plumbing failure, condensation, or air-carried moisture, which look identical inside and have entirely different causes, owners and repairs. Reads the evidence that separates them — when it appears, what the weather was doing, whether it is warm or cold, whether it is clean, whether it stops — then narrows the source and names what would confirm it. Trigger on "there is a leak", "water is coming in", "a stain appeared on the ceiling", "is this condensation or a leak", "where is this water coming from", "leak-triage", or water appearing anywhere with no agreed cause. Carries no dew points, no rainfall figures and no test standards; every value comes from the building's own data or the tester. It diagnoses rather than repairs, does not apportion liability, and does not replace the enclosure consultant, the plumber, or the Architect of Record. | Stars: 1 | Updated | Claude CodeCodex | Verified | Install |
| leed-trackerDavidCho1999/archaiflow-skillsThis skill should be used when the user wants to build, update, or score a LEED v4.1 BD+C New Construction scorecard for an architectural project — by walking the credit list, declaring pursuit status (Yes / Maybe / No / N/A) for each credit, and summing points to predict certification level (Certified / Silver / Gold / Platinum). Trigger on "LEED scorecard", "LEED v4.1 score", "score this project for LEED", "LEED tracker", or pasted project program with sustainability goals. Reads bundled credit-by-credit point matrix from USGBC LEED v4.1 BD+C and outputs a CSV scorecard plus a markdown summary in `docs/leed/`. Does NOT make LEED eligibility judgments — the user declares what they intend to pursue; this skill organizes and scores. | Stars: 1 | Updated | Claude CodeCodex | Verified | Install |
| lessons-learnedDavidCho1999/archaiflow-skillsKeeps a lessons-learned log in docs/practice/lessons.md: what happened, what it cost or saved, why it happened as stated, and what to do differently, each entry tagged by phase and topic so it can be found when the situation recurs. Trigger on "lessons learned", "post-mortem", "what went wrong", "remember this for next time", "add a lesson", "lessons-learned", the close of a project, or the start of one, when the right move is reading the log rather than writing it. Lessons are recorded as you state them: the skill never invents a lesson, softens one, or assigns blame beyond what you said. At project start or before a phase it reads back the entries whose tags match, which is the entire return on writing them. A lesson is one incident; recurring patterns become visible because the log accumulates, not because the skill generalizes. | Stars: 1 | Updated | Claude CodeCodex | Verified | Install |
| markup-readerDavidCho1999/archaiflow-skillsReads a marked-up drawing — a redline, a scanned sketch over a print, a PDF with comments, a photograph of a drawing somebody wrote on — and turns it into a numbered list of what each mark asks for, who it is for, and whether it can be acted on as drawn. Separates the marks that instruct from the marks that ask, and flags every mark whose meaning cannot be determined, because a guessed markup is worse than an unread one. Trigger on "read this redline", "they marked up the drawing", "what does this markup say", "turn these comments into a list", "markup-reader", or a marked drawing handed over. Carries no conventions: what a colour or a symbol means is the office's or the marker's, and it is asked rather than assumed. `drawing-diff` compares two issued revisions; this reads what somebody wrote on one. It does not replace the Architect of Record. | Stars: 1 | Updated | Claude CodeCodex | Verified | Install |
| meeting-notes-cleanerDavidCho1999/archaiflow-skillsThis skill should be used when the user provides raw architectural project meeting notes — transcripts, STT output, .vtt/.srt files, or jotted text — and needs structured action items. Trigger on 'clean up meeting notes', 'meeting summary', 'process this transcript', or pasted multi-speaker content covering project decisions, schedule, or scope. Generates a structured meeting markdown in `docs/meetings/` and merges new To-dos into the project's `docs/todos.md`. | Stars: 1 | Updated | Claude CodeCodex | Verified | Install |
| mep-reviewDavidCho1999/archaiflow-skillsReviews a mechanical, electrical, and plumbing drawing set against the architectural and structural drawings across eight coordination axes: shafts and risers, ceiling and plenum depth, equipment rooms, horizontal distribution, intake and exhaust, service entries, maintenance access, and document coordination. Trigger on "MEP review", "MEP coordination", "check the MEP set", "will the ducts fit", "plenum depth", "shaft coordination", "is there room for the equipment", or a shared MEP sheet with a fit question. The skill supplies no dimensions of its own: every clearance, depth, and area comes from the drawings or from you. Writes a dated review file and appends to a project index, with every unverified axis counted on the face of the review. Does not size equipment and does not replace the Engineer of Record. | Stars: 1 | Updated | Claude CodeCodex | Verified | Install |
| minutes-readerDavidCho1999/archaiflow-skillsReads minutes written by somebody else — a contractor, a client, a project manager, a consultant — for the thing that matters and gets skipped: what they recorded about you. Minutes become the record if nobody objects within the period the contract or the covering note sets, so an action assigned to you that you did not accept, or a decision recorded that was not made, hardens into fact by silence. Trigger on "read these minutes", "the contractor sent minutes", "do we need to correct these", "did we agree to that", "minutes-reader", or minutes received from anybody else. Carries no contract terms; the objection period comes from the agreement or the minutes themselves. `meeting-notes-cleaner` turns your own raw notes into actions; this reads somebody else's finished document. It does not replace the Architect of Record. | Stars: 1 | Updated | Claude CodeCodex | Verified | Install |
| mockup-reviewDavidCho1999/archaiflow-skillsReviews a built enclosure mockup, visual or performance, against the documents rather than against how it looks: establishes what this mockup was built to prove before anyone walks up to it, works the observation in a fixed order so the buildable-but-wrong conditions get seen, separates what the mockup demonstrated from what it merely did not contradict, and records what becomes the accepted standard for the rest of the work. Trigger on "mockup review", "reviewing the visual mockup", "performance mockup results", "what do we check at the mockup", "mockup punch list", "mockup-review", or a mockup photographed or described with a request for what to look at. Carries no tolerances, no test criteria and no acceptance thresholds; every one of those comes from the specification. It is not a test report, does not witness testing, does not accept or reject on the Architect's behalf, and does not replace the enclosure consultant or the Architect of Record. | Stars: 1 | Updated | Claude CodeCodex | Verified | Install |
| model-health-auditDavidCho1999/archaiflow-skillsAudits a building model for the practices that decide whether it stays usable: what the model is actually for and whether it is being built to that, whether geometry is modeled or drafted over, whether the data the project committed to deliver is present and consistent, what is warning or duplicated, and whether the file can be shared with consultants without an explanation. Reports by frequency and separates what will get worse from what is merely untidy. Trigger on "model health check", "audit this Revit model", "why is the model so slow", "is the model ready to share", "BIM standards audit", "model-health-audit", or a model handed over before an exchange. Carries no software-specific counts, no thresholds and no office BIM standard; the standard and the delivery requirements come from your BEP or your office. It reviews practice rather than drawings — `drawing-standards-audit` reads the sheets. It does not replace the BIM manager or the Architect of Record. | Stars: 1 | Updated | Claude CodeCodex | Verified | Install |
| nyc-zoning-lookupDavidCho1999/archaiflow-skillsThis skill should be used when the user asks NYC-specific zoning questions and wants citation-backed answers from the NYC Zoning Resolution. Trigger on NYC zoning terms (FAR, lot coverage, setback, side/rear yard, sky exposure plane, height factor, dwelling unit factor, use group, accessory use), district codes (R1-R10, C1-C8, M1-M3), Special Districts (Special Hudson Yards District, Special West Chelsea District, etc.), or direct §X-XX citation requests. Downloads the official NYC ZR Complete PDF (~100 MB) on first use, caches a section-to-page outline at ~/.archtmpl/pdfs/nyc-zoning/, and answers each query by reading targeted pages and citing them as `§X-XX, p.N`. Operates with zero external dependencies — no Python, no embeddings, no API key beyond Claude Code subscription. Do NOT trigger for non-NYC zoning (use zoning-summary) or NYC Building Code questions (different document). | Stars: 1 | Updated | Claude CodeCodex | Verified | Install |
| option-compareDavidCho1999/archaiflow-skillsBuilds a side-by-side comparison of design options in docs/design/options.md: the criteria you choose, one column per scheme, every cell filled from what you supply, and a tradeoff summary that stops short of picking the winner. Trigger on "compare options", "option A or B", "scheme comparison", "comparison matrix", "which scheme", "options for the client", "option-compare", or pasted facts about two or more schemes. Areas can be read from a space-program file where one exists, costs arrive from you or a cited benchmark, and empty cells stay visibly empty rather than estimated. Weighted scoring runs only if you supply the weights, arithmetic shown. The decision is the client's and yours; the skill structures it and does not make it. Renders to Word or PDF through aec-doc-gen when the meeting needs a handout. | Stars: 1 | Updated | Claude CodeCodex | Verified | Install |
| parti-studyDavidCho1999/archaiflow-skillsWorks the organizing idea of a scheme rather than its appearance: states the parti in one sentence, tests whether the plan actually obeys it, finds the places it breaks and asks whether each break is a compromise or a second idea competing with the first, and checks the parti against the constraints it has to survive. Generates alternatives only as different organizing ideas from the same constraints, never as variations of one. Trigger on "parti", "what is the organizing idea here", "is this scheme actually about anything", "concept diagram", "the plan feels arbitrary", "parti-study", or a scheme handed over with a question about what holds it together. Carries no formal vocabulary and no house style; the idea is yours and the skill tests whether the drawing keeps it. It does not critique the design — `design-critic` and `independent-critic` do that, deliberately separately — and does not replace the Architect of Record. | Stars: 1 | Updated | Claude CodeCodex | Verified | Install |
| passive-house-trackerDavidCho1999/archaiflow-skillsThis skill should be used when the user wants to track a Passive House project against the published certification thresholds — PHIUS+ 2021 (PHIUS, North American climate-specific targets) or PHI (Passive House Institute, fixed metric criteria — Classic / Plus / Premium). The user supplies modeled values (annual heating demand, annual cooling+dehumidification demand, source / primary energy, air leakage, peak loads, overheating frequency) from WUFI Passive or PHPP; the skill compares against published targets and emits a pass/fail-per-metric scorecard. Trigger on "PHIUS+ score", "Passive House tracker", "PHI Classic check", "compare WUFI/PHPP results to PHIUS+ targets", or pasted modeling outputs. Reads bundled PHIUS+ 2021 methodology + PHI fixed criteria from `references/` and outputs CSV + markdown summary in `docs/passive-house/`. Does NOT generate climate-specific PHIUS+ targets — user must fetch project-specific targets from the PHIUS portal (phius.org/phius-source-zero) before scoring. | Stars: 1 | Updated | Claude CodeCodex | Verified | Install |
| penetration-reviewDavidCho1999/archaiflow-skillsWorks through every penetration of a rated, acoustic or structural assembly and asks the three questions that get answered separately and should be answered together: what performance the assembly has to keep, what physically closes the hole, and whether a listed or tested system exists for that penetrant in that assembly. Builds the register across disciplines, since penetrations are made by trades who never see each other's drawings. Trigger on "penetration review", "firestopping check", "what closes this hole", "acoustic penetrations", "sleeve schedule review", "penetration-review", or an MEP set laid over the rated wall plan. Carries no listings, no ratings and no acoustic values; the required performance comes from your code analysis and every closure is a listed or tested system you cite. For the assembly's own continuity use `design-fire-barrier`; for air and water use `air-barrier-trace`. It does not replace the Architect of Record. | Stars: 1 | Updated | Claude CodeCodex | Verified | Install |
| permit-trackerDavidCho1999/archaiflow-skillsKeeps one dated log of every approval a project has to obtain, from the agencies you name rather than from an assumed list: what each permit is, which agency issues it, what it depends on, who owns it, what has been submitted and when, every comment received and its response, and every expiry. Reads back what is on the critical path and what is waiting on somebody. Trigger on "permit tracker", "track the approvals", "what permits do we need", "permit log", "who owns this approval", "when does the permit expire", "permit-tracker", or a list of agencies named with a request to keep track of them. Carries no agency lists, no review durations, no fee schedules and no permit types; every one comes from the jurisdiction, because they differ between adjacent municipalities and change without notice. It does not file anything, does not predict a duration, and does not replace the expediter, the code consultant, or the Architect of Record. | Stars: 1 | Updated | Claude CodeCodex | Verified | Install |
| phasing-studyDavidCho1999/archaiflow-skillsWorks out the phasing for construction in and around a building that stays occupied: which functions must never close, where each one lives during each phase, the swing space that makes the moves possible, and the sequence that gets everyone home without a circular dependency. Builds the occupancy ledger, the move list, and the per-phase checklist of what must be true before a phase can start. Trigger on "phasing study", "occupied renovation", "how do we renovate without closing", "swing space", "decant plan", "phasing-study", or a renovation scope paired with functions that cannot stop. The sequencing logic is the skill's; every duration, cost, and code determination comes from elsewhere: the contractor owns means, methods, and the schedule, and whether egress and separations hold during each phase is a code and AHJ question this skill raises and never answers. It is the design-side plan that makes those conversations possible, not a construction schedule and not a logistics plan. | Stars: 1 | Updated | Claude CodeCodex | Verified | Install |
| photo-readerDavidCho1999/archaiflow-skillsReads any photograph an architect ends up with — a site visit, an existing building, a defect somebody wants explained, a phone picture from a client, a product on a shelf — and separates what is visible from what is inferred, because a photograph is the most persuasive and least complete document on a project. Establishes when and where it was taken and which way the camera faced before reading anything, since a photograph with no location is evidence of nothing. Trigger on "read this photo", "what am I looking at here", "the client sent a picture", "is this a problem", "photo-reader", or an image handed over with a question. Carries no dimensions and measures nothing from an image. It diagnoses nothing: a photograph of a stain is not a diagnosis, and `leak-triage` takes it from there. It does not replace the Architect of Record. | Stars: 1 | Updated | Claude CodeCodex | Verified | Install |
| pipeline-trackerDavidCho1999/archaiflow-skillsKeeps one dated record of every pursuit the practice is chasing, from the first mention of a project — a call, a rumor, a client hint — through qualification, the go decision, the proposal, and the outcome, with the effort each pursuit consumed and the reason each loss happened. Leads every read-back with the pursuits consuming effort that have no go decision on record, because chasing without deciding is where business development money quietly goes. Trigger on "pipeline tracker", "what are we chasing", "track our pursuits", "proposals out right now", "did we ever hear back on that RFP", "what is our win rate", "pipeline-tracker", or a pile of pursuit emails handed over to make sense of. Carries no win-rate benchmarks, no hit-rate targets, and no probability weightings; the practice's history is the only pattern it reports. | Stars: 1 | Updated | Claude CodeCodex | Verified | Install |
| plan-check-response-builderDavidCho1999/archaiflow-skillsTurns a plan check or plan review comment set into a response matrix that answers every comment individually and on the record: splits comments that contain more than one question, sorts each into agreed, clarification, or disagreement, points every response at the specific sheet and detail that changed, and keeps the disagreements civil and evidenced because those are the ones that get read twice. Trigger on "plan check comments", "respond to the plan reviewer", "correction letter", "plan review response", "resubmittal response matrix", "plan-check-response-builder", or a set of reviewer comments pasted with a request to answer them. Carries no code positions of its own and never argues a clause from memory; every response either cites the drawing that changed or quotes the code text you supply. For a letter to the department rather than a comment matrix use `/draft-letter`; for tracking the submission itself use `permit-tracker`. | Stars: 1 | Updated | Claude CodeCodex | Verified | Install |
| plan-reviewDavidCho1999/archaiflow-skillsReads any floor plan you hand it and reports three things separately: what is drawn, what a plan of this kind should show and does not, and what the plan commits to without saying so. Works whatever the project type, whatever the phase, and whether the plan is yours, a consultant's, a client's previous architect's, or one found in a property listing. Trigger on "read this plan", "what do you make of this floor plan", "review this layout", "they sent us a plan", "plan-review", or a plan handed over with no particular question attached. Carries no dimensions and no standards; anything measured comes from what is on the drawing. `design-critic` judges a design, `code-compliance-review` audits against eight code axes, and `circulation-study` traces the journeys; this reads the drawing first, which is what the other three assume has happened. It does not replace the Architect of Record. | Stars: 1 | Updated | Claude CodeCodex | Verified | Install |
| post-occupancy-evaluationDavidCho1999/archaiflow-skillsGoes back to a finished building and finds out what actually happened: what people do that the design did not expect, what they have changed, what they work around, and what performs differently from what was intended. Separates observation from opinion and from measurement, keeps the visit away from the questions that only produce politeness, and ends by naming who each finding is for, since a finding nobody receives is a visit nobody should have made. Trigger on "post occupancy", "POE", "how is the building working", "go back and see how it is used", "soft landings", "post-occupancy-evaluation", or a building in use with a question about it. Carries no benchmarks and no satisfaction scales; targets come from what the project set and measurements from whoever takes them. `lessons-learned` records what a project taught the practice; this finds out what the building is doing. It does not replace the Architect of Record. | Stars: 1 | Updated | Claude CodeCodex | Verified | Install |
| precedent-critiqueDavidCho1999/archaiflow-skillsCritiques the user's scheme through lenses extracted from reference projects the user supplies, instead of a generic rubric. Trigger on "critique this through these precedents", "hold my scheme against these references", "critique this like these projects", "precedent-critique", or a scheme shared alongside two or more admired projects. The user supplies both sides: their own scheme (concept text, plans, sections, images) and material for each reference (project descriptions, photographs, drawings, published texts). The skill derives three to five lenses from the supplied material, confirms them with the user, then reads the scheme through each lens. It never characterizes a firm or a project from memory: a name without material stops the run. Divergence from the references is reported and questioned, never scored. It does not design, does not rank, and does not replace the designer's judgment or a design review. | Stars: 1 | Updated | Claude CodeCodex | Verified | Install |
| precedent-finderDavidCho1999/archaiflow-skillsThis skill should be used when the user describes a project concept, program, atmosphere, or site condition and needs architectural precedents. Trigger on 'find references', 'similar projects', 'precedents for', or natural-language descriptions paired with phrases like 'something with this feeling'. Caches a curated precedent list in `docs/precedents/` (one file per topic) and updates the project precedent index. | Stars: 1 | Updated | Claude CodeCodex | Verified | Install |
| program-readerDavidCho1999/archaiflow-skillsReads any brief or programme document — a client's schedule of accommodation, a spreadsheet from a user group, a consultant's brief, a previous architect's programme — and establishes what its numbers actually are before anybody designs against them: which are measured, which are standards, which are wishes, what area basis they are on, whether the total is the sum of the parts, and what the document counts that a building would not. Trigger on "read this brief", "they sent a schedule of accommodation", "what does this programme actually say", "is this area realistic", "program-reader", or a brief handed over. Carries no area standards and no benchmarks; the basis comes from the document or from `area-measurement`. `brief-review` interrogates a brief for gaps and contradictions; this reads one for what its figures rest on, and it runs first. It does not replace the Architect of Record. | Stars: 1 | Updated | Claude CodeCodex | Verified | Install |
| project-trackerDavidCho1999/archaiflow-skillsKeeps one dated log of a project's issues, decisions, risks, and action items in docs/project/tracker.md. Trigger on "project tracker", "decision log", "issue log", "risk register", "action items", "track this", "log this decision", "what did we decide", "who owns this", "project-tracker", or when meeting notes produce action items worth keeping past the meeting. Every entry carries a date and a person; closed items move to a closed section rather than being deleted, so the log is also the project's history. The skill never invents a status, an owner, or a deadline. What it was not told, it records as unassigned and says so. Feeds on meeting-notes-cleaner output and is what generate-weekly-report reads. Not a construction schedule and not a CA log; RFIs and submittals have their own logs. | Stars: 1 | Updated | Claude CodeCodex | Verified | Install |
| project-website-builderDavidCho1999/archaiflow-skillsTurns a project deck PDF, renders, site photos or drawing images into one web page that reads on a phone and can be sent as a link. Uses only the words and numbers in the files or typed in the chat, lists what it found and asks how each part should be shown (full width or grid, side by side or slider, one scroll or one section per screen) before building, keeps PDF drawings sharp when zoomed, and writes the page as one index.html that a drag-and-drop host can put online. Trigger on "make this a website", "project website", "send the deck as a link", "shareable project page", "a page the client can open on their phone", "project-website-builder", or files attached with a request for something a client can open. It does not write project copy, and it does not put anything online by itself. | Stars: 1 | Updated | Claude CodeCodex | Verified | Install |
| punch-listDavidCho1999/archaiflow-skillsMaintains a project punch list in docs/construction/punch-list.md. Trigger on "punch list", "punch item", "snag list", "deficiency list", "defect list", "punch walk", "punch-list", or a pasted site visit report whose observations should become tracked items. Each item carries a location, a description, the responsible trade, the date raised, its source, and a status that separates completed from verified, because a contractor marking work done is not the same as the architect seeing it done. The skill never invents an item, a responsible party, or a completion; what it was not told stays open. Feeds on create-site-visit-report output. It does not decide substantial completion and does not certify anything; those are contract acts. | Stars: 1 | Updated | Claude CodeCodex | Verified | Install |
| rfi-trackerDavidCho1999/archaiflow-skillsKeeps one dated RFI log and reads what the log says that no single RFI does: every RFI with who owns it, when it arrived, when the contract required an answer, when the answer went, and how it was classified, plus the running counts that reveal a pattern — a sheet generating a stream of questions, a contractor sending already-answered ones, a consultant consistently late, a set of clarifications quietly accumulating into a change. Trigger on "RFI log", "RFI tracker", "how many RFIs are open", "are we late on RFI responses", "which drawings are generating RFIs", "rfi-tracker", or an RFI register handed over to keep. Carries no response periods and no classifications of its own; the contract sets the turnaround and each RFI's class comes from whoever answered it. It does not answer RFIs — `/draft-rfi-response` does — and does not replace the Architect of Record. | Stars: 1 | Updated | Claude CodeCodex | Verified | Install |
| rfi-writerDavidCho1999/archaiflow-skillsThis skill should be used when the user provides an informal construction-site question — drawing/spec discrepancies, dimension conflicts, missing details — and needs a formal Request for Information (RFI) document filed in the project. Trigger on phrases like "make an RFI", "draft an RFI", "send this as an RFI", or pasted site questions implying transmittal to design consultants. Generates a numbered RFI markdown file in `docs/rfis/` and updates the project's RFI log. | Stars: 1 | Updated | Claude CodeCodex | Verified | Install |
| rfp-response-builderDavidCho1999/archaiflow-skillsBuilds an RFP or RFQ response around the evaluation criteria rather than around the practice's usual material: pulls every requirement and every question out of the document into a compliance matrix, follows the issuer's structure and numbering exactly, and checks before submission that every requirement has an answer and every page limit and format rule is met. Trigger on "respond to this RFP", "RFQ response", "build the proposal response", "do we answer everything in this RFP", "rfp-response-builder", or a solicitation handed over to answer. Carries no boilerplate and no standard firm material; everything comes from the practice's own content. For a federal qualifications package use `sf330-builder`; to decide whether to respond at all use `go-no-go-review` first. It does not decide the fee and does not replace the principal. | Stars: 1 | Updated | Claude CodeCodex | Verified | Install |