Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add pedroknigge/arkgate --skill ark-exploregit clone --depth 1 https://github.com/pedroknigge/arkgateWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/pedroknigge/arkgate/ark-explore)<a href="https://agentmods.dev/skills/pedroknigge/arkgate/ark-explore"><img src="https://agentmods.dev/badge/skills/pedroknigge/arkgate/ark-explore/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/pedroknigge/arkgate/ark-explore"><img src="https://agentmods.dev/badge/skills/pedroknigge/arkgate/ark-explore.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00024 | $0.09604 |
| Opus 5.5 | $0.00010 | $0.03842 |
| Sonnet 5.5 | $0.00005 | $0.01921 |
| Haiku 4.5 | $0.00002 | $0.00960 |
Grade A, and why
ark-explore scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 23d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 548 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/ark-explore — Recon the real project (map only)
Contener · Guiar · Ordenar. In plain words: contain the write, guide the next step, order leftover mess. This door is Guiar: map leftover design so you can order a messy tree. Do not apply. Skills never enforce — CLI / hooks / CI do.
When: you need a map, leftover design work remains after imports look green, or
one bounded decision needs 2–3 enforceable options.
Not when: the user wants edits applied (/ark-autopilot) or session 0 (/ark-adopt).
Steps
- Read
arkgate-check --doctor. - Open the real tree (≥12 files across ≥4 dirs). CLI JSON alone is incomplete.
- Rank leftover work. Hand off to
/ark-autopilotto apply one refactor. Do not write config or product code here.
Checklist
- Doctor #1: if ENFORCE + empty plan A and gates are already installed → Shape (
/ark-explorethen/ark-autopilot). Do not say install-agent-gates. - When
--require-gates/ adopted-strict is on and doctor JSON hasadrPresence, the next step is a short note underdocs/adr/(ordocs/decisions/). Soft. Not every change. Off when require-gates is off. No/ark-adr. WhenpolicyDeltaneeds--policy-ack(weaken, new layer, or new allow edge), the ack must name that note asadrPath— a reason alone is not enough. - When doctor JSON has
statesTransitionsorstatusTransitionCatalog, the next step is a short states → transitions table (or one link) on the domain doc already in play. Use catalog names when present. Soft. Silent when that doc is absent or Domain has no status vocabulary. No/ark-states. Do not invent states. - When doctor JSON has
noDomainFrontend, Domain is empty and the UI holds the rules. Next: one Domain file (/ark-place) then one small refactor (/ark-autopilot). Soft. Silent when there is no frontend or Domain already has files. No/ark-domain. Do not invent a Domain house. - When doctor JSON has
prototypeShortcuts, SQLite / a JSON-file store is standing in for Persistence, or admin / god-mode literals sit outside Domain (or an auth tag). Next: one Persistence adapter or Domain policy file (/ark-place) then one small refactor (/ark-autopilot). Soft. Silent when those houses are absent, the shortcut is already in the right house, or there are no such markers. No/ark-prototype. Do not invent a Persistence house. - When the tree is adopted and doctor JSON has
invariantTestsPath(or the check namesINVARIANT_TESTS_PATH_MISSING), the next step iscoverage.testGlobsorcoverage.coverageRootson a real tests folder. Fail-closed. Silent when not adopted, the catalog is empty, or every invariant sets coverage.test false. Handoff/ark-adoptto write the path. No new skill. - When any invariant is enforced and doctor JSON has
invariantCoverageRoots(or the check namesINVARIANT_COVERAGE_ROOTS_MISSING), the next step iscoverage.coverageRootson the folder the test runner uses. Fail-closed.testGlobsalone is not enough. Silent when no invariant is enforced. Handoff/ark-adoptto write the roots. No new skill. - When an invariant declares
coverage.symbol, read the witness path (symbolEvidenceFile) and the discard line on doctor and--rules-inventory, including when the run is green. A declaration in a non-test file is not “the tests pass” — ArkGate never runs tests. Depth and byte caps stay internal; the counts are what you read. - When doctor JSON has
orphanModuleswith files listed, map them: nothing imports each file and no entry point covers it. Readcertainty—maybe-dynamicandmaybe-unresolvedmay still be loaded (open what the evidence names). Apartialsection is never a complete list.--doctor --alladds exports nothing imports by name. Map only; deleting is/ark-fix, one file at a time. A framework or loader ArkGate does not know goes in.ark/entry-points.jsonwith a reason. Never say "dead code". No new skill. - A doctor green / Healthy line must name a file, config key, or test.
Uncited green is a lie — treat it as unfinished. No
/ark-cite. - Distinguish missing skills vs stale catalog. Installed ≠ stale.
- Hook green is not tree green (scripted edits bypass PreToolUse).
- Bug / false green / false red / missing doc / improvable behavior in ArkGate itself → draft one upstream GitHub issue (
pedroknigge/arkgateor the installed packagerepository.url), ask the human to confirm send, thengh issue create. Never auto-file. Never file ArkGate defects on the consumer product repo.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 23d ago Changed · +7 lines 5034ab311271
- 27d ago Changed · +2 lines 99783b396908
- 28d ago Changed · +12 lines 486a270bc8a2
- 1mo ago Changed · +8 lines 123a6f35f7dd
- 1mo ago Changed · +5 lines 7ca98fd435cb
- 1mo ago Changed · +8 lines e572e5bf487b
- 1mo ago Changed · +4 lines 62e36f0add89
- 1mo ago Changed · +8 lines d758a9fd98b7
- 1mo ago Changed · +1 lines 51987d112220
- 1mo ago First seen · 493 lines · 24 tokens per session scan A f1970d7d3bfa
ark-explore is a skill published in the GitHub repository pedroknigge/arkgate (1 stars, last pushed today), licensed MIT. It adds 24 tokens to every session and 9,604 once invoked, about $0.0001 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
architecture-patterns
Software architecture patterns and best practices.
Architecture Change Communication
Ensure architecture updates document affected stakeholders, notification plan, and deprecation/retirement signals to keep knowledge aligned.
code-review-architecture
Architecture-focused code review covering hexagonal boundary violations, DDD anti-patterns, CQRS misuse, and microservices coupling issues. Applied in addition to the language-specific review skill when architecture markers are detected. Invoked when reviewing hexagonal architectures, DDD patterns, or microservices…
design-patterns
Design Patterns, SOLID & Clean Architecture: Reference guide and advisor for software design patterns, SOLID principles, Clean Code practices, and Clean Architecture. Use whenever the user asks about design patterns, SOLID principles, Clean Code, Clean Architecture, DDD, refactoring, code organization, coupling…
improve-codebase-architecture
Use when surfacing architectural friction inside a single EVOKORE bounded context and proposing deepening refactors (shallow modules, leaky seams, low locality) that turn shallow modules into deep ones — informed by ADR-0005 bounded contexts and the project's domain language.
architecture-patterns
Software architecture patterns and best practices.