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 kardebadas/claude-plugin --skill fast-pipelinegit clone --depth 1 https://github.com/kardebadas/claude-pluginWrote 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/kardebadas/claude-plugin/fast-pipeline)<a href="https://agentmods.dev/skills/kardebadas/claude-plugin/fast-pipeline"><img src="https://agentmods.dev/badge/skills/kardebadas/claude-plugin/fast-pipeline/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/kardebadas/claude-plugin/fast-pipeline"><img src="https://agentmods.dev/badge/skills/kardebadas/claude-plugin/fast-pipeline.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 | $0.00095 | $0.01616 |
| Opus 5.5 | $0.00038 | $0.00646 |
| Sonnet 5.5 | $0.00019 | $0.00323 |
| Haiku 4.5 | $0.00010 | $0.00162 |
Grade A, and why
fast-pipeline 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 17d 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 — 114 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Fast Pipeline
Overview
Request → approved spec → plan → implementation → review/fix loop, with one
human gate (the spec) and no project tests written. It composes
Superpowers skills plus superb:bug-fix and adds nothing else.
The user's invocation of this skill is a standing instruction. Where a sub-skill's default says otherwise, the overrides below win — user instructions outrank skills. Violating the letter of an override to follow a sub-skill default is violating the user.
The no-tests rule
No new unit, integration or end-to-end tests. No TDD steps, no test tasks, no
"add a regression test", no test-fix loops. Verification is a fast check only:
build, typecheck, lint or syntax check, whichever the repo has. An existing
suite runs only where a sub-skill runs it once as a gate (e.g.
finishing-a-development-branch); a failure there is reported, not chased
with new tests.
| Rationalization | Reality |
|---|---|
| "writing-plans' template requires a failing test first" | The user waived tests. The template is a default; drop its test steps. |
| "TDD is mandatory regardless of preference" | Not in this run. The user tests manually. |
| "bug-fix is not done without a regression test" | Take bug-fix's own "cannot be covered by a test" branch, citing this rule. |
| "The reviewer flagged missing tests" | Not a finding here. Drop it. |
Stages
- Explore. Targeted, read-only. Independent areas → REQUIRED:
superpowers:dispatching-parallel-agents(one Explore agent per area). Return: where it belongs, relevant files, patterns/APIs to reuse, similar existing code, integration points, risks. Sequential areas stay one agent. - Ask. Only what changes behaviour, scope, architecture, UX, data
handling or a major decision, and is not answered by the request, the
conversation, the repo or its conventions. One batched
AskUserQuestion, recommendation first. None needed → ask none. - Spec. REQUIRED:
superpowers:brainstorming, fed the request, answers and exploration findings. Overrides: exploration and questions are done (do not repeat them); always write the spec file, even for a bounded change; recommend one approach; skip the visual companion; present the whole spec once, not section by section. Content: behaviour, scope, non-goals, affected components, interfaces, data flow, edge cases, constraints, acceptance criteria. No filler. - Gate: user approves the spec. Changes requested → revise through brainstorming, re-present. Nothing past this point until approved.
- Pressure-test the spec. Dispatch a fresh general-purpose agent
(never the author) with the spec path, the request and the exploration
summary. It reads the repo and returns contradictions, missing
requirements, unsupported assumptions, wrong beliefs about the code,
integration conflicts, key edge cases and needless complexity — each
tagged
material(changes approved behaviour or scope) orminor. Fixminorin the spec directly; takematerialto the user. No re-approval of minor fixes. - Plan. REQUIRED:
superpowers:writing-plans. Overrides: no test steps or test tasks; Global Constraints carriesNo new tests — the user tests manually. Verify with: <fast check command>; each task ends with that check then commit; Review Focus lines get no tests — they are handed to the stage 9 reviewer; no docs or cleanup tasks the spec did not ask for. The execution method is pre-chosen — subagent-driven — and the plan is not sent to the user for review. Commit the spec and plan. - Pressure-test the plan. Another fresh agent checks the plan
against the codebase: paths exist or belong, named functions and
interfaces are real, ordering and dependencies hold, nothing duplicates
existing code or fights project patterns, the spec is fully covered, no
extra work. Fix the plan directly; go to the user only for a
materialchange. Then continue at once. - Implement. REQUIRED:
superpowers:subagent-driven-development. Keep its mandatory per-task reviews; add no other. Its final whole-branch review is stage 9, and its single fix wave is replaced by stage 10. After stage 10 is clean, resume its Finish (the rulings list, thenfinishing-a-development-branch). - Final review. REQUIRED:
superpowers:requesting-code-reviewover the whole branch (merge-base..HEAD).PLAN_OR_REQUIREMENTS= spec path + plan path (its Review Focus lines are what to probe) + "No tests by user decision; do not report missing tests." Findings that count: bugs, spec violations, bad integration, regressions, security, data loss, broken error handling, false assumptions, serious maintainability. Cosmetic and out-of-scope refactors do not. - Fix loop. Any Critical or Important → REQUIRED:
superb:bug-fix, one invocation per round with every such finding, each reviewerfile:lineas its report (repro:manual — user tests). Overrides: no regression test (see the no-tests rule); work in the current worktree and branch; its executor skips its own final whole-branch review andfinishing-a-development-branch— this run re-reviews and finishes once. Then run stage 9 again. Repeat until no Critical or Important remain. Minors are listed in the final report, never looped on. A finding that survives three rounds goes to the user.
What ships with it
1 file 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.
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.
- 17d ago First seen · 114 lines · 95 tokens per session scan A 339a78a62a5f
fast-pipeline is a skill published in the GitHub repository kardebadas/claude-plugin (3 stars, last pushed 18d ago), licensed Apache-2.0. It adds 95 tokens to every session and 1,616 once invoked, about $0.0004 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 2026-09-28.
Other skills, from other repositories
import
Import a self-contained NeatContext context bundle shared by another person, or reconcile a newer copy of a context already on this machine, leaving the source bundle unchanged. Use only when the user explicitly invokes this skill or asks to import a NeatContext bundle.
save
Save durable decisions, findings, plans, and implementation knowledge from the visible Codex conversation into a new or existing NeatContext context. Use only when the user explicitly invokes this skill or asks to preserve the current conversation as reusable context.
create
Create a fresh NeatContext context from a user-defined behavioral profile and an existing local knowledge folder. Use only when the user explicitly invokes this skill or asks to create a new context rather than save the current conversation.
extensions
Report which extensions the connected NeatContext context expects, and whether this machine provides them. Use when the user asks what a context can reach, why an extension tool is missing, or how to connect one here.
delete
Preview and permanently delete a NeatContext context while preserving any externally linked knowledge folder. Use only when the user explicitly invokes this skill or clearly asks to delete a context.
export
Export a NeatContext context saved from a conversation into a self-contained bundle folder that can be shared or moved to another machine. Use only when the user explicitly invokes this skill or asks to export a NeatContext context.