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 agentmods add skills/imodoiepale/unleashed-loop.dev-skill/loop-apinpx skills add imodoiepale/unleashed-loop.dev-skill --skill loop-apigit clone --depth 1 https://github.com/imodoiepale/unleashed-loop.dev-skillWhat 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 | $0.00190 | $0.02819 |
| Opus 5 | $0.00095 | $0.01409 |
| Sonnet 5 | $0.00038 | $0.00564 |
| Haiku 4.5 | $0.00019 | $0.00282 |
Grade A, and why
loop-api scanned grade A with 1 finding 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 yesterday.
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.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
and repository classes, don't hand them a `curl` command and a global function. How it starts
The opening of the file, as written. The whole thing — 232 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Loop API
Help developers integrate with the Loop developer API — and be right, because this is a banking API where a wrong endpoint or a misread error code costs real money and real time.
The one rule that matters
Every factual claim you make about Loop — a URL, a header, a field name, an error
code, a limit — must come from a file you have just read in references/.
Do not answer from memory or from general knowledge of "how banking APIs usually
work". Loop is a specific product; plausible-sounding invented endpoints are the main
way an assistant wastes a developer's afternoon. references/ is generated directly
from Loop's published documentation by tools/ingest_docs.py, so every line in it
traces back to a source URL and a fetch date recorded at the top of the file.
If the answer is not in references/, say so plainly and point the developer at the
relevant doc page. "I don't see this documented — here's the page to check" is a
genuinely useful answer. A confident guess is not.
Start here
Read references/INDEX.md first. It lists every documentation page with its topic, so
you can load only the one or two files that bear on the question instead of pulling
the whole corpus into context.
Then read the specific file(s) you need. Common entry points:
references/api-flows.md— what each operation actually does, end to end: the order the gateway runs its checks, the point at which money moves, and the retry decision tree. Read this for any "is this possible?" or "why did this fail?" question — it also carries a table mapping common product ideas to documented / partly-supported / undocumented, which is the fastest honest answer to a feasibility question.references/conventions.md— the cross-cutting request/response rules. Read this before writing any integration code, not just when asked about it. It carries the rules a developer is most likely to get wrong in a way that costs money rather than merely failing.references/signing.md— if the product requires signed requests, an implementation that is subtly wrong fails in ways that look like an auth problem. It carries published test vectors; have the developer verify against those before debugging against the live API.references/doc-conflicts.md— check this before telling a developer their code is wrong. Where the vendor's own documentation contradicts itself, a request that "matches the docs" is ambiguous, and this file is often the fastest explanation for an inexplicable failure.references/coverage.md— check this before saying a capability doesn't exist. Absent from the corpus is not the same as absent from the product.references/manifest.json— what was captured, from where, when, and by which capture method. Check thefetcheddate when a developer reports that reality disagrees with the docs, and thecapturefield when precision matters.- Authentication pages — read these before answering any question that involves a request, because nearly every failure a developer brings you is an auth failure wearing a costume.
- Endpoint pages — for request/response shapes and required fields.
What ships with it
22 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- references/api-flows.md 11 KB
- references/authorisation.md 4.6 KB
- references/conventions.md 7.1 KB
- references/coverage.md 4.9 KB
- references/doc-conflicts.md 8.1 KB
- references/INDEX.md 5.1 KB
- references/introduction.md 2.3 KB
- references/loop-prompt.md 6.8 KB
- references/manifest.json 3.4 KB
- references/overview.md 4.2 KB
- references/pay-to-loop-till.md 5.1 KB
- references/pay-to-mpesa-paybill.md 5.7 KB
- references/pay-to-mpesa-till.md 5.3 KB
- references/README.md 2.4 KB
- references/security-testing.md 11 KB
- references/send-money-loop.md 6.7 KB
- references/send-money-mpesa.md 8.3 KB
- references/send-money-pesalink.md 6.2 KB
- references/signing.md 5.0 KB
- references/transaction-status-inquiry.md 6.0 KB
- scripts/search_docs.py 3.4 KB runs code
- scripts/validate_skill.py 6.7 KB runs code
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.
- yesterday First seen · 232 lines · 190 tokens per session scan A 09bb39dda532
loop-api is a skill published in the GitHub repository imodoiepale/unleashed-loop.dev-skill (4 stars, last pushed 19d ago), licensed MIT. It adds 190 tokens to every session and 2,819 once invoked, about $0.0010 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
next-cache-components-adoption
Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…
babysit-pr
Babysit a GitHub pull request after creation by continuously polling review comments, CI checks/workflow runs, and mergeability state until the PR is merged/closed or user help is required. Diagnose failures, retry likely flaky failures up to 3 times, auto-fix/push branch-related issues when appropriate, and keep…
imagegen
Generate or edit raster images when the task benefits from AI-created bitmap visuals such as photos, illustrations, textures, sprites, mockups, or transparent-background cutouts. Use when Codex should create a brand-new image, transform an existing image, or derive visual variants from references, and the output…
cpu-profile-analysis
Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…
next-cache-components-optimizer
Drive a Next.js route to instant navigation by setting up an agentic loop, under Cache Components / PPR, on initial load (hard navigation) and client-side navigation (soft navigation). Encode the goal as a failing @next/playwright instant() e2e and work it to green, one verified route at a time; the shipped test then…