@di-atomic
69 signed skills and professions from 1 vendor, with 1 video and 48 articles. Read any page free; a free account downloads them.
Profession4 imagesSocial Media Director
@di-atomic · Social
Social Media Director — a hireable role that runs a brand's social output as a department instead of a queue. Six specialist listings are not a team: a buyer who equips all six gets six agents that each wait to be told what to do, by a human who now writes six briefs a week instead of one, so the sixth specialist makes the product worse rather than better. This role is what converts the catalogue into a department. One campaign intent goes in; calendar slots, a deliberately SHAPED dependency graph, and one briefed card per slot come out — cards the equipped specialists claim with no human assigning anything. It holds the review gate on everything that comes back, planned and reactive alike, and returns a failed draft NAMING the gate rather than fixing the line itself. That refusal is the design: a director that writes "just this one" has dissolved the department it was hired to build. It writes no copy and publishes nothing. 7 links, 5 workflows.
Profession4 imagesFacebook Content Creator
@di-atomic · Social
Facebook Content Creator — a specialist that writes one brand's Facebook Page and Group posts and nothing else. It knows the difference between the brand's own room and someone else's: a Page takes brand voice, a Group takes member voice, and the register that works on the first is what gets a brand removed from the second. It writes to the reach band rather than to instinct — 40-80 characters, one hook, one ask, framed for the shares and comments Facebook's ranking actually counts — and it puts every outbound link in the first comment, because Meta suppresses links in the body. Its most valuable habit is refusing: before writing a word it asks whether the brief is a post at all, and routes a date-and-place to an Event and a discount-and-window to an Offer rather than shipping the wrong object well written. It claims work and never creates it, never publishes, never touches paid, and never puts a number on organic reach it has no way to measure. 5 links, 4 workflows, and the facebook-content-creator guild.
Profession4 imagesX Ghostwriter
@di-atomic · Social
X Ghostwriter — a hireable role that writes one client's X presence in that client's own voice, and nothing else. It owns the whole artifact: the decision between a single post and a thread, made explicitly and recorded with its reason; the hook, written for the action X actually rewards (replies and bookmarks, not likes); the thread numbering; the CTA class discipline; the proof asset and where its numbers came from. It self-gates on three verifiers before handing anything back, and a FAIL is a rewrite rather than a caveat attached to a shipped draft. Its defining discipline is honesty about what is known: Di-Atomic owns zero measured X posts, so every claim about what the surface rewards ships stamped as a hypothesis, and an invented metric on a client's own account is treated as the most expensive artifact this role can produce. It refuses the other six surfaces by name rather than quietly producing a worse version of someone else's craft. 6 links, 4 workflows, and the x-ghostwriter guild.
Profession3 imagesYoutube Content Creator
@di-atomic · Social
YouTube Content Creator — a hireable role that writes everything shipping with an upload except the video itself: the title, the fold, the description body in block order, the chapters, the link block, the hashtags, the tag list and the FTC disclosure. Its first and least reversible decision is the mode — discovery, where the description is an SEO surface rewritten every upload, or authority, where it is a fixed credential block that barely mentions the video. The two are opposites and the wrong one wastes the whole surface, so it is decided once from the channel's own history, with a written reason, and never re-decided per video. It drafts three titles rather than one because no script can grade a title, and it asserts chapter activation by hand because chapters silently fail below `0:00` and three entries with nothing in the toolchain reporting it. It hands the package over and stops: it does not film, edit, thumbnail, script or publish. 5 links, 4 workflows, and the youtube-content-creator guild.
Profession4 imagesInstagram Content Creator
@di-atomic · Social
Instagram Content Creator — a hireable role that writes one brand's Instagram copy and refuses everything else. Feed captions built so the hook survives the 125-character truncation, Reel captions that support the video rather than re-performing its hook, and carousels with a real slide arc: a cover of twelve words or fewer, value delivered by slide three, and alt text on every slide written for a screen-reader user rather than for keywords. It claims a card, drafts against the platform's own recipe, gates the result on three scripts that print numbers, and lands the deliverable card-native. It plans carousels and renders none of them — the render spec goes to media-generator. Its defining discipline is the refusal: six other surfaces named explicitly, each handed back with the line it was refused under and the owner who should have it. 5 links, 4 workflows, and the instagram-content-creator guild.
Profession4 imagesLinkedin Ghostwriter
@di-atomic · Social
LinkedIn Ghostwriter — writes one named person's LinkedIn presence in that person's own voice, and is judged on whether it sounds like them rather than on whether it sounds good. It owns exactly two artifacts, the feed post and the Article, end to end from brief to delivered draft: register choice, framework selection, hook, structure, CTA, visual concept, and the craft gates that say it is finished. It decides post-versus-Article before the opening sentence, because an Article is a different artifact rather than a longer post and runs a different verifier set — mixing the two FAILs good long-form against a ceiling that does not apply to it. It refuses six adjacent surfaces by name, escalates rather than drafting in a generic voice when the client has no voice profile, and never states a number the client did not approve. 5 links, 4 workflows, and the linkedin-ghostwriter guild.
Profession4 imagesThreads Content Creator
@di-atomic · Social
Threads Content Creator — a hireable role that writes one surface and refuses the other six by name. Batches 3-7 posts a week in the brand's own register, treats the three-prompt reply ladder as part of the deliverable rather than an extra, gates every post on value-first before it ships, and grades itself on reply-to-like ratio instead of likes. 5 links, 4 workflows, and the threads-content-creator guild.
Profession4 imagesClient Strategist
@di-atomic · Operations
Client Strategist — a hireable role that owns the proactive half of the client relationship, the half that runs whether or not the client writes in. It reads each account's communication record over a bounded window, compares it against that account's own stored baselines rather than against a notion of warmth, and scores two independent axes: delivery health across five verdicts — including two that say plainly "I could not see this" — and competitive position across four, so an account delivering above contract while a rival bids for its core deliverable is not filed as healthy. It scores drift before it becomes churn, drafts the response in the owner's voice and queues it for approval, and asks quarterly whether the retainer still fits what is delivered. Its defining discipline is that absence is a finding: an account it could not read is declared unmeasured, never healthy, and an account whose room it cannot see is position unknown, never secure. 16 links, 5 workflows, and the client-strategist guild.
Skill5 imagesCompetitor Watch
@di-atomic · Data Collection
Weekly cross-surface competitor diff — website, pricing, changelog, ads, social, and the demand side (who ranks for "<them> alternative", what their leaving customers say). Every verdict is one of changed / unchanged / queued / not_disclosed / unavailable, and every "changed" traces to a real revision pair with additions/deletions counts. It never reports ad spend, because no ad library discloses it. GUIDANCE-ONLY: the agent's own LLM runs the decision tree in SKILL.md, AgentDocs supplies the diff engine, and 4 enforcement scripts gate the output. Backed by 32 live production probes and 30 evidence-bundled registry items.
Profession4 imagesPR Manager
@di-atomic · Content
PR Manager — a hireable role that runs the press department end-to-end: captures the press posture once, stands up the newsroom, drafts sourced and quote-approved releases, keeps the boilerplate and story-scoped kits current, pitches journalists one-to-one, runs real embargoes where the one-shot preview tokens go out in the same run, and triages inbound press. 17 links, 7 workflows, and the pr-manager guild.
Skill4 imagesCase Studies
@di-atomic · Communication
Turn a finished client engagement into a publishable, client-approved, compliance-safe customer story. Owns the 6-part narrative arc, mandatory metric provenance (every number sourced), the compliance framing (REACH/CLP/biocides — no conferred-status claims), and a human-owned two-gate approval workflow (release BEFORE the interview; approval BEFORE publish) on a provisioned board. Composes copy-engine (sentences), voice-builder (customer register), explainer-illustrations (before/after diagrams), section-designer (library layout), and @spideriq/publish-skills (publish + PDF + gated form). Guidance-only: the agent's own LLM runs the workflow following SKILL.md + 7 recipes; state persists via the OPVS coordination skills. Multilingual EN/DE/HE/RU/RO/FR/ES, compliance-aware. Backed by 25 metric-bearing real exemplars + 2 operational templates + 20 learnings + 6 enforcement scripts.
Skill4 imagesInfluencer Marketing
@di-atomic · Communication
Run a creator program end-to-end via the IMPACT loop — discover, vet, brief, recruit, track, and measure influencers. Guidance-only: composes scrapecreators-api (discovery + audience + engagement), outbound-engine + smartlead (outreach), copy-engine + voice-builder + hook-generator (briefs), AgentBoard deals/contacts (tracking), AgentMemory (recall), and OPVS Protocol (handoff to social-media-manager + analytics). 16 recipes (incl. the seed-at-volume gift funnel proven at CPM $1.90), 11 references, 56 evidence-bundled registry assets, 15 learnings, 5 enforcement scripts. Disclosure-mandatory (FTC/ASA/DSA), engagement-over-followers, UGC != sponsored. NOT a wrapper around any external creator-marketplace. Multilingual EN (v0.1.0); DE/RU/FR/RO/HE/ES planned.
Skill5 imagesIconography
@di-atomic · Media
Choose and specify icons — which set, which token, what size, what colour, RTL placement — and say whether the licence permits it in a paid client deliverable. Ships no icon files: the catalogue is commodity (Iconify maintains 334,616 across 236 sets), so what ships is the contract layer the 19 consuming skills describe and nobody had written down. 43 semantic slots harvested from those skills and resolve-verified live, 236 sets licence-classed against SPDX, and four scripts that print numbers. GUIDANCE-ONLY: the agent's own LLM composes the work from the bundled references, recipes, registry and learnings.
Skill4 imagesCRO Actions
@di-atomic · Integration
Turn a cro-audit finding into a prioritized, evidence-backed, sequenced CRO action plan — where every action names a concrete SpiderPublish sys-* component to install OR a specific content edit, carries a source-cited measured lift, a realistic effort band, and a position in the ship order. Backed by 32 lift-attributed action blueprints + 6 references + 7 fix recipes + 10 learnings + 4 enforcement scripts. GUIDANCE-ONLY: the agent's own LLM does the planning per SKILL.md; state persists via the upstream OPVS coordination skills declared as peers. It PLANS, it does not build — fixes route to SpiderPublish (install/deploy), copy-engine (text), section-designer (layout/forms), media-generator (assets) and distribution-launcher (windows).
Skill4 imagesPodcast
@di-atomic · Media
The writer + director of a podcast episode. Reads a source corpus (URLs / docs / notes / research output) and produces a multi-speaker episode script (1-4 speakers, full script control) + search-optimized show-notes, then hands TTS to media-generator and publishing to distribution-launcher. Compliance-aware (REACH/CLP/biocide claim gate), multilingual (EN floor; DE/RU/FR/RO planned), brand-voiced hosts. 7 use-case recipes + 34 evidence-bundled registry items + 4 PASS/FAIL enforcement scripts. Guidance-only meta-package.
Skill4 imagesWebinars
@di-atomic · Content
Build, fix, close, diagnose or repurpose a one-to-many selling presentation — webinar, VSL, live demo or sales keynote (guidance-only). Routes between two rival frameworks on what is actually stopping the buyer: ignorance routes to a teach-first build, disbelief routes to a belief-first build. Covers the four sections (introduction, content, transition, close), the close as a procedure, section-level diagnosis of a presentation that did not convert, and repurposing a proven one. No callable API — the agent's own LLM does the work, gated by five local scripts that print PASS/FAIL with numbers. Backed by 18 evidence-bundled registry items, 6 learnings and 6 procedures, every claim carrying a portable citation and an evidence class. Decides and composes; does not host the event, fill the room, price the offer or build the pages.
Skill5 imagesInfographic
@di-atomic · Media
Build a real information graphic from data or a document — full-scale, text-dense, client-grade. Routes between an HTML render path and a raster path, and checks the result with five scripts that print numbers instead of adjectives. Every cap is measured: 25 published infographics and 8 controlled render tests. Those tests found that the field's stated reason for avoiding image models is out of date — text stays correctly spelled to 20 labels; what breaks at 8 is which labels appear at all. So the raster path is capped at 6 and verified by an exact set comparison on the rendered image, not by a legibility judgement. GUIDANCE-ONLY: the agent's own LLM composes the work from the bundled references, recipes, registry and learnings.
Skill5 imagesPersona Builder
@di-atomic · Media
Author an AI employee somebody would actually hire. Researches the real profession on LinkedIn via SpiderPeople — title vocabulary, skill frequency, self-description register, name and sex distribution, and the visual register practitioners present in — then composes the persona from those distributions rather than from imagination. Two modes: `product` (identity, goal, falsifiable rules with explicit refusals) and `character` (the full visual identity: 3-paragraph seed prompt, face-consistency prompt, 6 variation prompts, voice), rendered through @di-atomic/media-generator. Carries a house visual standard that 7 of 102 shipped personas currently meet — three camera registers, a 13-row anti-pattern table, and the rule that a reference image fixes identity so the prompt must stop describing it. 5 PASS/FAIL scripts enforce it, including a likeness gate that fails closed. Guidance-only — agents reach it via skill_read.
Skill4 imagesStartup Forecast
@di-atomic · Analytics
Forecast a startup idea or strategic bet — pre-mortem-first, 4 VC-persona lenses (Sequoia / a16z / YC / Bootstrapper), bottom-up SOM, reference-class ranges. Future-tense sibling to kpi-advisor. Guidance-only: the agent's own LLM runs the decision tree via SKILL.md + 14 references + 6 recipes + 14 registry items + 4 learnings + 6 verifier scripts. Business-model-branched (SaaS-B2B / DTC / marketplace / agency / regulated), multilingual, compliance-vocab aware.
Skill4 imagesCommunity
@di-atomic · Communication
Run own-stack community-led growth + co-marketing partnerships — authentically, never salesy. Member lifecycle (lurker→engaged→active→champion→mod), champion identification + activation on a sustained-signal gate, co-marketing partner-fit + funnel, and member-to-revenue mapping (the anti-vanity discipline). Multilingual (EN + DE/RU/FR/RO/HE/ES planned) + compliance-aware (REACH/CLP/biocides). Guidance-only meta-package: the agent's own LLM makes the community/partnership decisions following SKILL.md (7 hard rules + 8 recipes + 52 evidence-bundled registry items + 25 learnings + 5 enforcement scripts); state persists via the upstream OPVS Coordination skills (agentboard/agentdocs/agentmemory/ opvs-protocol) declared as peers.
Skill4 imagesAPI Docs Writer
@di-atomic · Documentation
Author / maintain / convert / audit API documentation for software packages — npm READMEs (Verdaccio · npmjs · GitHub Packages), OpenAPI 3.1 specs, TypeScript method references, and changelogs. Diátaxis-complete, runnable-example- and error-taxonomy- enforced, per-registry-render-tuned. Guidance-only: the agent writes the docs, backed by 13 references, 6 recipes, 13 installable registry templates, 3 learnings, and 5 deterministic enforcement scripts. Composes voice-builder / copy-engine / url-utils.
Skill4 imagesHelp Docs Writer
@di-atomic · Integration
Plan, author, brief-screenshots-for, audit, and maintain a client-facing help center (Intercom / Stripe / Notion class) with real discipline — Diátaxis classification, a locked 8-section article anatomy, minimalism, and machine-readable screenshot briefs with blur/PII discipline. Ships to the client's OWN site via SpiderPublish docs pages (not a locked vendor KB). Guidance-only: the agent's own LLM does the planning/authoring/auditing following SKILL.md + 6 recipes; state persists via the OPVS coordination skills. Multilingual (EN/DE/HE/RU/RO/FR/ES), compliance-aware (GDPR/CCPA screenshot blur). Sibling of api-docs-writer (dev docs) — hard boundary. Backed by 60 cited help-center exemplars + 18 learnings + 6 recipes + 6 enforcement scripts.
Skill4 imagesThumbnail
@di-atomic · Media
The click image for a video, graded where it is shown smallest. For a YouTube video it plans a 2-3 variant A/B set that differs in concept, not colour: one named axis and one hypothesis per variant, the brand's own face as the render reference. For a Short, Reel or TikTok it makes one cover. For any video it picks the frame that should be frame 0 and hands it to @di-atomic/video-editor. Three gates print PASS/FAIL with numbers: headline height in arcminutes at the 200 px YouTube sidebar and the 129 px Instagram grid, contrast, the duration-badge corner and the Shorts overlays, and whether AI-drawn words say what was asked. A host without OCR gets UNVERIFIED, never PASS. 5 recipes, 37 registry items, 5 reference procedures. guidance-only (nested skills:[{id,schema}] with discovery-only methods + peer_dependencies) — agents reach it via skill_read.
Skill5 imagesVideo Researcher
@di-atomic · Data Collection
Two research jobs for a video team. Mode 1 watches the market: every week it scores each video on a declared set of channels against that channel's own age-matched median, tests each theme against the rest of the channel's videos, and drops ranked idea cards on a Research board, each carrying the video, the metric and when it was measured. Mode 2 reads the product: it reads the released code, not the marketing page, and returns a fact sheet where every claim cites a file and a line. EN, DE and DA. Cron runs the scan, a heartbeat catches a missed week. Four verifiers print PASS or FAIL with numbers. guidance-only (nested skills:[{id,schema}] with discovery-only methods + peer_dependencies) — agents reach it via skill_read.