Claude

Conducts blameless postmortems that transform failures into learning opportunities by documenting timelines, quantifying impact, performing root cause analysis (5 Whys, fishbone diagrams), and defining corrective actions with owners and deadlines. Use when analyzing failures, outages, incidents, or negative outcomes, conducting blameless postmortems, identifying corrective actions, learning from near-misses, establishing prevention strategies, or when user mentions postmortem, incident review, failure analysis, RCA, lessons learned, or after-action review.

Checks every post currently assigned to a substacker section against that section's promise and flags posts that no longer fit. Distinguishes acceptable-stretch (minor) from borderline (surface for review) from genuine-drift (violates promise). Never reassigns automatically — only flags. Use on every Curator run where at least one section already exists. Trigger keywords — drift, drift audit, section fit, promise violation, post in wrong section.

Matches the kind of debt a firm issues to the assets it finances — maturity, currency, fixed or floating, and features such as convertibility or commodity linkage — using asset duration and the sensitivity of firm value and operating income to interest rates, inflation, real growth and exchange rates. Use when advising on what debt to issue, reviewing an existing debt structure, choosing debt maturity or debt currency, deciding fixed versus floating, or after finding an optimal debt ratio.

Analyzes decisions from multiple stakeholder perspectives (engineering, product, legal, finance, users) to uncover blind spots, surface tensions, and synthesize alignment paths with explicit tradeoffs. Use when stakeholders have conflicting priorities, need to pressure-test proposals, build cross-functional empathy, or when user mentions "what would X think", "stakeholder alignment", "see from their perspective", "blind spots", or "conflicting interests".

Scans a substacker draft line-by-line against the canonical voice-profile.md don't-list and signature moves. Emits phrase-level flags with location, quoted phrase, violation type, voice-profile citation, and up-to-2 suggested rewrites per flag. Use as pass-2 skill (voice) after structural-review completes, when a draft reads competent but not in the writer's voice, or when the writer asks "does this sound like me?" Trigger keywords — voice check, delve, unpack, paradigm shift, sounds AI, does this sound like me, voice violation.

The note-writing discipline for this vault's evergreen knowledge graph, modeled on a Zettelkasten reading companion and governed by the vault conventions. Enforces declarative-claim titles, one claim per note (atomicity), own-words prose with no block quotes, the piped [[slug|Title]] link form, the labeled link-relationship vocabulary (Confirms/Contradicts/Extends/Context/Prerequisite/Builds-on/Applies/Example-of/Contrasts-with), 3-6 links per note, and search-before-create deduplication. Use when capturing a claim as an evergreen note, refactoring a sprawling note into atomic ones, or wiring a new claim into the graph. Trigger keywords — evergreen note, zettel, atomic note, declarative title, link relationships, search before create, route the claim.

Builds interactive D3 visualizations of genetics and genomics concepts for the vault's docs/ GitHub Pages site — each one a teaching object with a takeaway-stating title, an annotation layer, and a guided reading order. Specializes d3-visualization, visual-storytelling-design, and cognitive-design for the crop-genetics / genomic-selection domain, supplying a concept-to-visualization catalog, preattentive encoding guidance, colorblind-safe palettes, accessibility requirements, and a fixed docs/ file layout. Delegates design review to the cognitive-design-architect agent and runs design-evaluation-audit and cognitive-fallacies-guard before publish. Use when turning an evergreen claim into an interactive figure, building or revising a docs/ visualization, or planning the site's viz layout. Trigger keywords — D3 viz, interactive figure, GitHub Pages chart, Hardy-Weinberg, Manhattan plot, allele drift, LD decay, breeder's equation, genomic prediction accuracy, reaction norm, kinship heatmap.

Decomposes high-level North Star metrics into actionable sub-metrics and leading indicators, maps causal relationships between metric levels, and identifies high-impact experiments to move key metrics. Use when setting product North Star metrics, decomposing business metrics into drivers, mapping strategy to measurable outcomes, identifying which metrics to move through experimentation, understanding leading vs lagging indicators, prioritizing metric improvement opportunities, or when user mentions metric tree, metric decomposition, North Star metric, KPI breakdown, metric drivers, or how metrics connect.

Assigns 1-4 topic tags to a seed body from the controlled vocabulary in substacker shared-context/topic-ledger.md. Prevents tag sprawl at small-corpus scale by requiring existing-tag match or logged addition. Uses keyword + title match; logs near-miss candidates to pending-tags. Use after format normalization and before dedupe. Trigger keywords — tag, topics, categorize, classify, taxonomy, controlled vocabulary, topic ledger.

Computes one player's expected FIFA World Cup Fantasy points for a single round (xEV) from a scout signal — folding start probability, the 60-minute appearance cliff, fixture-and-minutes-scaled npxG/xA, clean-sheet and defensive-action floors, the penalty/set-piece premium, and the card/concede downside into one number plus its full decomposition, variance, ceiling, and floor. This is the atomic EV unit the whole system samples from. Use whenever a single player's per-round point distribution is needed — called by wc-scout, every wc-strategist genotype, wc-matchday-tactician, and wc-fitness-eval (which sums these into raw_xEV). Reads the scout signal; never re-scouts. Emits a player-ev signal.

A strict advisory-only editing discipline for a writer who dictates ("speaks out") essays and wants help WITHOUT having their voice changed. The editor directs structure, flags grammar, and suggests strategic language — but never modifies the writer's text unless the writer explicitly says "apply" / "make that change" / "rewrite this." Produces a line-referenced, suggestion-only critique where every item is marked the writer's call. Four passes — structural, line (grammar/clarity), voice, pre-publish. Use when reviewing a draft, critiquing a spoken-out article, or doing a pre-publish check. The companion to learning-in-public-voice and the operating manual for the biostat-editor agent.

Creates concise, decision-ready product specifications (one-pagers and PRDs) that align stakeholders on problem, solution, users, success metrics, and constraints. Use when proposing new features/products, documenting product requirements, creating concise specs for stakeholder alignment, pitching initiatives, scoping projects before detailed design, capturing user stories and success metrics, or when user mentions one-pager, PRD, product spec, feature proposal, product requirements, or brief.

Answers "what have I already thought about X?" by searching the substacker corpus (seeds, drafts, published) for seeds matching a topic, keyword, analogy, or author. Returns a ranked list of seeds with id, title, status, density score, and a one-line excerpt. Use when another agent (Intuition Builder, Editor) needs prior thinking before generating new material, or when the writer asks "have I written about X." Trigger keywords — search, find, what have I, already thought, prior work, precedent, have I written about.

Transforms documents containing theoretical knowledge or frameworks (PDFs, markdown, book notes, research papers, methodology guides) into actionable, reusable Claude Code skills using systematic reading methodology. Use when user mentions "create a skill from this document", "turn this into a skill", "extract a skill from this file", or when analyzing documents with methodologies, frameworks, or processes that could be made actionable.

Scans a substacker draft for 10 signatures of AI-generated explainer slop — meta-framing openers ("In this post"), list-heavy argument, nominalization clusters, generic examples lacking first-person texture, prompt-residue phrases ("Let's break this down"), buzzword stuffing, outline-shaped paragraphs, hedge clusters, flattened uncertainty. Use when a draft "feels generic" even after voice-check passes. Trigger keywords — slop, AI-written, generic, template, meta-framing, zombie nouns, prompt residue, outline-shaped.

Guides cooking through culinary principles, food science, and flavor architecture rather than rote recipe steps. Covers technique teaching (knife skills, sauces, searing, braising), food science (Maillard reaction, emulsions, brining), flavor troubleshooting (salt/acid/fat/heat balance), menu planning, ingredient substitutions, plating, and cultural cuisine exploration. Use when users mention cooking, recipes, chef, cuisine, flavor, technique, plating, food science, seasoning, or culinary questions.

Assigns each extracted claim to one of five buckets — simplified-correct, simplified-boundary, wrong, contested, overclaim — with low/medium/high confidence and one-sentence rationale. Classification happens before primary-source verification (which confirms, not invents). Use for every claim from claim-extractor. Trigger keywords — classify, bucket, simplified vs wrong, claim type, technical classification.

Calculates probability-weighted averages of all possible outcomes to enable rational decisions under uncertainty. Covers scenario identification, probability estimation, payoff quantification, and risk-adjusted interpretation. Use when comparing risky options (investments, product bets, strategic choices), prioritizing projects by expected return, assessing whether to take a gamble, or when user mentions expected value, EV calculation, risk-adjusted return, probability-weighted outcomes, or decision tree.

Classifies every hedge in a substacker draft as either a precision hedge (keep — "n=1 may not replicate", "I do not know") or an epistemic-weakness hedge (flag — "I think", "perhaps", "arguably", "it could be argued"). Only flags weakness hedges; suggests either a commit (remove hedge, take position) or a specific hedge (name the uncertainty). Use when a draft feels wishy-washy or when a cluster of modal verbs appears. Trigger keywords — hedging, I think, perhaps, arguably, uncertainty, weak claim, wishy-washy.

Generates 3-5 candidate first-line hooks for a specific platform (Substack Note, X, LinkedIn, cross-post) from a given spine. Uses platform-appropriate hook patterns (confession / claim / question / reframe) and voice-profile constraints. Runs before each platform rewrite so the rewrite skill picks the strongest hook rather than reusing the essay's opener verbatim on every platform. Trigger keywords — hook, opening line, platform hook, first tweet, LinkedIn hook, Substack Note hook.

Computes FAAB (Free Agent Acquisition Budget) recommended and maximum bids for Yahoo fantasy baseball waiver targets. Implements the baseball-specific layering of the faab-bid-framework (positional_need_fit, role_certainty, urgency, season_pace, league-inflation calibration) and DELEGATES the game-theoretic primitives -- first-price shading and winner's-curse haircut -- to the sibling skills `auction-first-price-shading` and `auction-winners-curse-haircut`. Produces a recommended bid, a hard ceiling, a rationale with the full delegation chain, and guardrail flags. Use when the user asks "how much should I bid on X", mentions FAAB bid, waiver bid amount, blind bid, Yahoo waiver claim sizing, or when mlb-waiver-analyst needs a bid amount for an identified target.

Scores the deployment leverage of all five FIFA World Cup Fantasy boosters (Wildcard, 12th Man, Maximum Captain, Clean Sheet Shield, Qualification Booster) across the remaining tournament horizon — computing, for each chip and each candidate round, the EXTRA expected points the chip yields in that round versus an average round, then turning the leverage surface into a fire/hold call per chip this round plus a maintained earmark table (chip -> earmarked round -> trigger -> fallback). This is the math engine behind wc-chip-strategist. Reads the round's player-ev, clean-sheet, fixture/progression, ownership and captain-ladder signals plus tournament-state (chips remaining, phase, surviving nations); never re-derives them. Enforces the golden rule — never reach the semis with chips unused. Use whenever a chip decision is live (the chip-check prompt, the group->KO transition, or any round where a chip is earmarked or leverage spikes). Emits a chip-plan signal.

Extracts atomic technical claims from a substacker essay draft, converting flowing intuition-first prose into a numbered list where each item is a statement that could in principle be verified or falsified. Skips non-technical sections (personal anecdote, motivation, call-to-action). Use when the Technical Reviewer starts a per-draft review. Trigger keywords — extract claims, atomic claims, technical claim list, fact-check prep.

Evaluates the final paragraph of a substacker draft for compression and closing form — bolded maxim, forward-looking question, or compressed mechanism statement. For series posts (frontmatter series — {slug}), verifies the running scoreboard (P&L, Brier, W-L) is present and updated. Use on every draft. Blocks publication of series posts missing the scoreboard. Trigger keywords — closer, closing, last paragraph, bolded maxim, scoreboard, CTA, wrap up, conclusion.

Compares multiple named alternatives against weighted criteria to produce transparent, defensible choices with explicit trade-off analysis. Covers criterion identification, weighting approaches (direct allocation, pairwise comparison, stakeholder averaging), scoring calibration, sensitivity analysis, and group decision facilitation. Use when choosing between vendors/tools/strategies, balancing competing priorities (cost vs quality vs speed), or when user mentions "which option should we choose", "compare alternatives", "evaluate vendors", or "trade-offs".

Renders a markdown report to a PDF using pandoc with xelatex (11pt serif body, 1-inch margins, numbered footnotes, formal heading hierarchy). Requires a one-time install of pandoc and a LaTeX engine on the user's machine — basictex on macOS or texlive-xetex on Linux. Does not attempt automatic install. Fails loudly with the exact install commands if pandoc or xelatex is missing on the user's PATH. Use when producing a finished strategist or analyst report PDF from a polished markdown source.

Evaluates the first 1-3 sentences of a substacker draft against the writer's signature opener patterns — confession / "I hadn't done X" / reframe / small concrete admission. Classifies opener as confession | reframe | admission | news-hook | generic-opener and flags news-hook/generic as tier-1. The opener sets the voice contract for the essay. Use on every draft. Trigger keywords — opener, hook, first sentence, opening, confession opener, news hook, generic opener.

Converts one candidate cluster from cluster-corpus-by-theme into a named, promised section proposal ready for writer review. Calls write-section-promise for the one-sentence promise. Rates fit confidence (high / medium / low / provisional) and flags borderline posts. Use once per cluster that passes ≥3-post threshold. Trigger keywords — propose section, section proposal, new section candidate.

Recommends structural cleanups for the substacker section map — sections to retire, sections to merge, posts to reassign. Applies under-filled, stale, and overlapping heuristics. Writes proposals with reasons-to-reject (steelman counter). Does not execute. Use once per Curator run, after drift audit. Trigger keywords — prune, retire section, merge sections, reassign post, cleanup.

Scores a FIFA World Cup Fantasy candidate squad/plan as risk-adjusted, rank-relative fitness under a stated rank objective (protect / gain / neutral). Sums fixture-scaled player xEV + captaincy uplift, adds clean-sheet correlation and progression-carry, applies an ownership-leverage term (cover chalk when protecting, fade it when gaining), and a variance term whose SIGN flips with the objective (penalty when protecting, reward when chasing) — the selection-pressure dial. Returns one fitness number plus the full decomposition. Use to score the population in the evolution engine, or to evaluate any single squad/plan. The objective makes the same candidate score differently for a leader vs a chaser; never returns raw expected points.

Documents significant architectural and technical decisions with full context, alternatives considered, trade-offs analyzed, and consequences understood. Creates a decision trail that helps teams understand why decisions were made. Use when choosing between technology options, making infrastructure decisions, establishing standards, migrating systems, or when user mentions ADR, architecture decision, technical decision record, or decision documentation.

Grounds visual design decisions in cognitive psychology principles — perception, attention, memory, Gestalt grouping, and visual encoding hierarchy — explaining WHY certain designs work. Covers interfaces, data visualizations, educational content, and presentations. Invoke when user mentions cognitive load, visual hierarchy, working memory, preattentive processing, Gestalt principles, encoding hierarchy, or cognitive design pyramid. For design evaluation, use `design-evaluation-audit`. For fallacy prevention, use `cognitive-fallacies-guard`. For data storytelling, use `visual-storytelling-design`.

Computes substacker's rolling 4-week baseline for open rate, click rate, views-per-send, and weekly subscriber delta using corpus/stats/ archived CSVs. Produces per-metric z-scores of the current week against the baseline and flags cold-start windows where fewer than 4 prior weeks exist. Use after ingest-substack-csv each Monday. Trigger keywords — baseline, rolling median, z-score, cold start, per-metric comparison.

Guides creation of custom, interactive data visualizations with D3.js — bar/line/scatter charts, network diagrams, geographic maps, hierarchies, and real-time data updates with zoom/pan/brush interactions and animated transitions. Use when chart libraries (Highcharts, Chart.js) lack the customization needed and you require low-level control over data-driven DOM manipulation, scales, shapes, and layouts. Invoke when user mentions D3, d3.js, custom visualization, force-directed graph, or data-driven SVG.

Decomposes complex unknowns into estimable components to produce rapid order-of-magnitude answers with bounded uncertainty. Use when making quick estimates (market sizing, resource planning, feasibility checks), bounding unknowns with upper/lower limits, sanity-checking strategic assumptions, or when user mentions Fermi estimation, back-of-envelope calculation, order of magnitude, ballpark estimate, or triangulation.

Normalizes a single inbox file of any supported format (plain markdown, Claude.ai JSON export, Claude Code JSONL session, Readwise markdown/CSV highlight, transcript with timestamps or speaker labels, link capture) into a clean markdown body plus partial frontmatter (id, title, source block, word_count). Handles format-specific failure modes — JSON content-block arrays, timestamp stripping, per-highlight chunking, URL-vs-commentary separation. Use when ingesting any inbox item for the substacker Librarian. Trigger keywords — normalize, convert, parse, transcript, export, JSON, JSONL, highlight, CSV.

Scores and ranks substacker Trend Scout annotated candidates against voice-profile and goals, producing a top-10 keep list and an explicit drop list with reasons. Weighted-sum scoring across intuition-density fit, goal alignment, dedup penalty, source reliability, freshness. Produces the digest's keeps and drops sections. Use after cross-ref-topic-ledger. Trigger keywords — rank, fit score, user fit, keep list, drop list, signal weight.

Plans backward from a fixed goal or deadline to the present, identifying required milestones, dependencies, critical path, and feasibility constraints to transform aspirational targets into actionable sequenced plans. Use when planning with fixed deadlines, working backward from future goals, mapping critical path, or when user mentions "backcast", "work backward from", "reverse planning", "we need to launch by", "target date is", or "what needs to happen to reach".

Provides the house style for analyst-grade strategist writing — third-person register with sparing first-person, no em dashes, no "not X, not Y, not Z" negation cascades, numbered footnote citations rather than inline source parentheticals, specific opinion-signaling phrases, and topic-forward paragraph structure modeled on voice patterns observed in Damodaran's Musings on Markets and Thompson's Stratechery. Use when consolidating working notes into a finished long-form strategist or analyst report that must read as written by a senior human analyst rather than an AI assistant.

Given a candidate item from the substacker Trend Scout fetch, WebFetches the full post or arXiv abstract and produces a one-line "teaches X" summary plus signal_type classification (mechanism / empirical / tool / opinion / announcement / benchmark). Distinguishes teaching-content from capability-announcement explicitly. Use during the weekly run, after fetching and before ranking. Trigger keywords — summarize, signal type, mechanism vs announcement, teaching content.

Proposes adds and removes to the Trend Scout watchlist based on consecutive-failure sources, repeated-reference external authors, and user-added feedback markers. Emits a proposed diff at ops/trend-scout/watchlist-proposed-diff.md for user review. On explicit approval, applies the diff. Monthly cadence. Trigger keywords — watchlist update, add source, remove source, watchlist review, monthly review, source pruning.

Applies a systematic three-pass revision system (Zinsser, King, Pinker, Clark) to existing drafts — Pass 1 cuts clutter, Pass 2 reduces cognitive load, Pass 3 improves rhythm. Use when revising, editing, or polishing drafts, cutting word count, tightening prose, improving readability, or fixing flow, or when user mentions revision, editing, cut clutter, too wordy, improve readability, fix the flow, reduce word count.

Rewrites a published substacker essay as three X thread variants (short 3-5 tweets, medium 6-8, long 9-12). Each tweet ≤280 chars. Hook tweet works standalone. No numbering by default (2026 convention for tech-first-principles accounts). Final tweet is the link. If essay doesn't translate to X, emits a VERDICT line and halts rather than producing weak variants. Trigger keywords — X thread, Twitter thread, thread, tweet, threaded post, thread variants.

Maintains a ledger of HSA-qualified medical expenses paid out of pocket (not reimbursed from the HSA), each with date, amount, provider, description, and a path to the scanned receipt. Tracks the running unreimbursed total — the amount of HSA balance the household can pull tax-free at any future date — and validates each candidate against IRS-qualified-expense categories. Use when a new medical receipt arrives, when totaling future tax-free HSA reimbursements, when planning a deferred reimbursement, or when user mentions HSA receipt vault, qualified medical expenses, or HSA shoebox strategy.

Ingests a single file from the substacker inbox/ into corpus/seeds/ as a normalized markdown seed with full frontmatter. Orchestrates format normalization, topic tagging, intuition-density scoring, dedupe, changelog, ledger update, and inbox-file move to .processed/. Use when the user drops raw material into inbox/ and runs /ingest, at session start, or whenever a single inbox file needs to become an indexed seed. Trigger keywords — ingest, inbox, new note, new transcript, new highlight, index this, add to corpus.

Structures thinking across multiple abstraction levels (30,000 ft strategic, 3,000 ft tactical, 300 ft operational) while maintaining consistency between layers. Guides through top-down decomposition, bottom-up aggregation, cross-layer translation, and constraint propagation. Use when reasoning across multiple abstraction levels, designing systems with hierarchical layers, explaining concepts at different depths, maintaining consistency between principles and implementation, or when users mention 30,000-foot view, layered thinking, abstraction levels, top-down design, or strategy-to-execution alignment.

Synthesizes 13 weeks of substacker Growth Analyst reports + the most recent Curator review + a meta-scan of the published corpus into a 400-700 word narrative that names the quarter's shape — what happened, what changed, what held steady, what surprised. Used by the Growth Strategist at the opening of every review. Trigger keywords — quarterly, zoomout, quarter narrative, rollup, what happened this quarter.

Scores prose with Flesch Reading Ease, Flesch-Kincaid Grade, SMOG, Gunning Fog, and Dale-Chall, reports every score on one shared grade-band scale, and rewrites the sentences that fail. Runs on markdown files or on piped draft text before it is sent, and can run automatically through Claude Code hooks. Use before delivering any substantial written output, when a draft reads as convoluted or bloated, when a document must hit a reading-level target, or when checking a folder of documents for prose quality. Trigger keywords — readability, reading level, Flesch, Flesch-Kincaid, SMOG, Gunning Fog, Dale-Chall, grade level, hard to read, convoluted, dense prose, plain language, reading ease.

Performs pass-1 structural review of a substacker essay draft — argument flow, out-of-order moves, buried topic sentences, missing pivots, weak signposting, paragraph-logic issues. Emits the "Argument flow" and "Structural blockers" sections of the Editor artifact. Use when reviewing a draft's macro-structure before addressing voice, when a draft feels like it meanders, or when the user asks whether the argument lands. Trigger keywords — structure, argument flow, outline, signposting, meandering, pivot, macro edit, substantive edit.

Identifies substacker seeds older than 30 days with status=seed and no incoming related_seeds links, flags them for writer review, and recommends keep / promote-to-draft / kill based on density score. Does NOT auto-execute any action. Emits a review list to ops/librarian/YYYY-MM-DD-stale-sweep.md. Run at session start after ingest, once per day max. Trigger keywords — stale, sweep, review, old seeds, cleanup, gardener, corpus hygiene.

Computes the optimal rolling-captaincy LADDER across a matchday's staggered kickoff sequence — the ordered captain candidates, the explicit bank-or-switch rules at each kickoff, and the captaincy_uplift that wc-fitness-eval folds into raw_xEV. Because the armband doubles the round's points AND can be MOVED during the round to a player who has not yet kicked off, captaincy EV is the expected value of the best outcome reachable by switching across the sequence — not a naive 2x of the top xEV player. Call from wc-matchday-tactician (to write MB2 of the matchday plan), from the strategists (to value a candidate's captaincy when building a matchday plan), and from wc-fitness-eval (which consumes captaincy_uplift). Returns the ladder, the switch thresholds, and the uplift against a stated baseline; the ladder is moot under the Maximum Captain chip (value = max of candidates).

Renders the FIFA World Cup Fantasy decision board — the advisory output that surfaces 2-4 genuinely distinct, weighted options for the expert manager and ends on "your call." Takes the evolution engine's verified offspring (with fitness decomposition, genotype lineage, and specialist dissent) and formats each as an option with its concrete picks, EV/ownership-leverage/variance/progression decomposition, football reasoning, what it's betting on, the dissent, and an ownership read — then a trade-off axis, a recommended-but-overridable default tied to the rank objective, and a "what I need from you." Enforces options-not-commands, shows the working, surfaces dissent, never auto-commits. Use to present any squad/matchday/transfer/chip decision.

Models the FIFA World Cup Fantasy field — effective ownership, the template set, the differential list, and the rank-protect-vs-gain leverage math — so the system reasons about "what will most managers do?" rather than raw points. Computes effective ownership EO = field_ownership% x (1 + captaincy_share), classifies every owned/missed player into the three regimes (own high-EO haul = hold station, MISS high-EO haul = drop hard, own low-EO haul = leap the field), builds the template set and differential list, flags where the field is over-concentrated (fade ops) and under-reacting (get-there-first), and returns the ownership_leverage contribution under the rank objective θ that wc-fitness-eval folds into fitness. Also computes mini-league rival-relative leverage. Emits an `ownership` signal. Use when called by wc-ownership-analyst, by the strategists weighting picks by leverage, and by wc-fitness-eval scoring the rank term.

Validates and persists FIFA World Cup Fantasy signal files to signals/YYYY-MM-DD-<type>.md. Checks the required frontmatter (type, round, date, emitted_by, confidence, source_urls), range-checks declared numeric signals, confirms every factual claim carries a source URL or "manager-provided", rejects unknown signal types, and refuses to persist a signal that fails validation (logging the failure instead). Keeps the inter-agent signal layer auditable so downstream agents can trust what they read and never re-derive it. Use whenever an agent or skill writes a signal.

Writes a 60-140 word third-person blurb for the Substack cross-post feature, positioned so another newsletter writer can paste it into their cross-post popup without editing. Third-person throughout ("In this piece, Kushal argues…"). No subscriber CTAs. Use as the cross-post arm of the Distribution Translator. Trigger keywords — cross-post, cross-poster, blurb, Substack cross-post, third person, positioning.

Designs structured scoring tools with explicit criteria, performance scales, and descriptors for consistent, transparent quality assessment. Use when need quality criteria and scoring scales to evaluate work consistently, compare alternatives objectively, set acceptance thresholds, reduce subjective bias, or when user mentions rubric, scoring criteria, quality standards, evaluation framework, inter-rater reliability, or grading/assessing work.

Fetches arXiv papers submitted within a given date window matching a keyword set, with optional restriction to one or more arXiv categories (e.g. cs.LG, cs.CL, cs.CV, stat.ML, math.ST, q-bio.QM). Wraps the public `export.arxiv.org/api/query` endpoint, parses Atom XML, handles pagination, and normalizes records to the same canonical shape as `fetch-preprint-recent` and `fetch-pubmed-recent` so a calling agent can dedupe across sources. Domain-neutral — usable for any literature scan that crosses CS, ML, statistics, math, physics, or quantitative biology. Use when user mentions arXiv, cs.LG, ML papers, NeurIPS-adjacent preprints, weekly arXiv scan, or when a literature-scan agent needs arXiv records alongside bioRxiv / medRxiv / PubMed.

Stress-tests predictions by assuming failure and working backward to identify blind spots, tail risks, and overconfidence. Applies Gary Klein's premortem technique to probabilistic forecasting. Use when confidence is high (>80% or <20%), need to identify tail risks and unknown unknowns, want to widen overconfident intervals, or when user mentions premortem, backcasting, what could go wrong, stress test, or black swans.

Names what the substacker writer should stop doing — habits with no evidence of use, goals that became theatre, sections with 2 consecutive dormant quarters, and agents in the team whose output the writer ignores. Produces a bulleted list, each item with one sentence of why. Max 4 items. Ordered by ease (easiest first). Used once per Growth Strategist review. Trigger keywords — kill list, stop doing, what to cut, dead habits, dormant, ignored output.

A closed-book-first probing protocol that assesses how well the learner actually holds a module's claims — asking before telling, probing the gap, detecting misconceptions, then confirming. Tags each probe with a Bloom level, places the learner in a Dreyfus mastery band (unfamiliar/aware/functional/proficient/fluent), scores the session, and schedules the next spaced-repetition surfacing on expanding intervals (1d, 3d, 7d, 16d, 35d) — pulling the next date earlier on a miss. Writes an assessment-session note to assessments/log/ and proposes a review-due update on each evergreen note touched. Use when checking mastery of a module, running a spaced-repetition review, or gating phase progression. Trigger keywords — assess mastery, review-due, spaced repetition, Bloom check, mastery band, am I ready to move on.