5dive-ai

Grade a delivered claim against the artifact instead of against the report of it, emitting a three-state verdict — pass, fail, or not-reached — with what you did not check stated explicitly. Use when verifying someone's finished work, confirming a fix actually landed, or self-auditing your own claim before you publish it.

End-to-end test and click-verify a web app with Playwright — write specs, drive authenticated pages, screenshot, and run in CI. Use when a web change has to be proven in a real browser rather than by unit tests.

Author your own OpenAgent persona file and mint its shareable holo trading card — write, validate, tier and render `<id>.persona.yaml`. Use when creating or updating YOUR identity as a persona or card, or provisioning a live teammate from one.

Edit drafts into sharper, more human writing while preserving the writer's personal voice, or detect AI-slop patterns without rewriting. Use when a draft should read clearer, more direct and less AI-sounding, or when asked whether writing reads as AI.

Author or modify a LOOP.md — the portable agenticloops.dev format for a recurring, scheduled agent (trigger + skills + prompt in one file). Use when writing a new loop or editing an existing one; to find and install a loop that already exists use `find-loops`.

Search the agenticloops.dev directory for an existing agentic loop (a recurring, scheduled agent) and install it. Use when someone wants a repeating job run on a timer and one may already exist; to author a new LOOP.md use `loops`.

Diagnose a failing 5dive agent runtime on its own VM — auth state, service health, disk, memory, recent CLI errors, skill integrity — and return a root-cause report, plus an optional host security audit. Use only when the 5dive runtime itself is suspect; a command, build or test failing inside your own work is not a trigger.

Run structured multi-source research that produces a cited, adversarially verified synthesis with explicit confidence. Use when a question needs more than a single lookup and being wrong is costly — comparing options, validating a claim, surveying a market, briefing a decision.

Write, rewrite or improve conversion-focused marketing copy for a web page — homepage, landing, pricing, feature or product: headlines, CTAs, value propositions, hero sections. Use when the page text itself has to persuade or convert; for email sequences see `emails`.

Compile durable, judgement-shaped knowledge into interlinked markdown — atomic files, [[wiki-links]], a maintained index. Use when persistence is asked for explicitly ("save this", "write it to the wiki") or when the work produced a decision record, a cause, or a gap analysis; ordinary factual answers and routine task results are not a trigger.

Review a diff or pull request for correctness first and quality second — logic bugs, edge cases, error handling, races, missing tests, API misuse, performance and readability. Use when reviewing someone else's PR, auditing a change before merge, or self-reviewing your own diff.

Generate and iterate paid-ad creative at scale — headlines, descriptions, primary text and full ad variations for any advertising platform. Use when producing or refining the ad copy itself; for campaign strategy and targeting see `ads`, for page copy see `copywriting`.

Administer the local 5dive runtime beyond everyday delegation: accounts and auth recovery, declarative fleets and remote boxes, every seat's live screen at once, loops and goal DAGs, governance, memory and wiki writes, plugins and host remediation. Use when runtime administration itself is the task; `5dive-cli` already covers spawning agents, messaging them and the shared task queue.

Operate the local 5dive runtime: coordinate sibling agents and inspect or update shared tasks. Use for delegation, inter-agent messaging and task-queue operations; use `5dive-cli-extras` for runtime administration beyond those.