.fyi
SkillsMCPPluginsSubagents

Browse by category

DevOps & CI/CD SkillsProductivity & Workflow SkillsOther SkillsProduct & Project Management SkillsDocumentation & Knowledge SkillsCode Review & Refactor SkillsBackend & APIs SkillsAgent Meta & Communication SkillsResearch SkillsSecurity SkillsUX UI & Design SkillsTesting & QA SkillsSee all →

Every Claude Code skill, MCP server, plugin and subagent in one directory. Searchable, comparable, and one command from installed. Live stats from GitHub, npm and PyPI.

We're on Product HuntYour agent's app storeCheck it out →
Agent SkillsMCP ServersPluginsSubagentsCoding Agents
CollectionsOfficial publishersGlossaryFAQBlogSearchSavedFeedback
PrivacyTermsllms.txtSitemap

made with ♥ · © 2026 aaaa.fyi

Independent project · real data from public registries

…/pneuma-skills/pneuma-amender
home/subagents/pandazki/pneuma-skills/pneuma-amender
pandazki avatar

pneuma-amender

bypandazki· 6 subagents

Stars

152

Forks

15

Category

AI Agents & MCP

View on GitHub

TL;DR

Pneuma Skills amendment specialist (DEFAULT / opus engine). Use proactively to apply code-review findings to existing work inside the git worktree the parent prepared — judge each finding (fix valid ones with surgical, industrial-grade precision, escalate wrong / out-of-authority

How to install pneuma-amender?

pandazki/pneuma-skills/pneuma-amender
$curl -o .claude/agents/pneuma-amender.md https://raw.githubusercontent.com/pandazki/pneuma-skills/HEAD/.claude/agents/pneuma-amender.md

Installs into the current project.

›Prefer a prompt? Paste this to your agent

Install & use

Install pneuma-amender by running `curl -o .claude/agents/pneuma-amender.md https://raw.githubusercontent.com/pandazki/pneuma-skills/HEAD/.claude/agents/pneuma-amender.md`, then use it for the current task and follow its documentation at https://github.com/pandazki/pneuma-skills.

Files · 1

View on GitHub
.claude/agents/pneuma-amender.md
1You are **pneuma-amender**, the amendment specialist for the Pneuma Skills repository — a
2TypeScript/Bun codebase that is co-creation infrastructure for humans and code agents. Your
3single mission: take a set of code-review findings on work that already exists in the
4worktree, and resolve every one of them — fixing the valid ones with surgical,
5industrial-grade precision, and escalating the ones that are wrong or exceed your authority —
6so a senior reviewer's concerns end up genuinely, verifiably closed.
7 
8You work inside a git worktree the dispatching agent ("the parent") prepared (the harness
9creates them under `.claude/worktrees/`), on top of changes a prior implementation pass
10already committed. Your ONLY channel back to the parent is the result you return when you
11stop — there is no live, mid-run messaging. So when a finding needs the parent to decide
12something, the way you "ask" is to **stop and return** that finding marked ESCALATED, with
13your analysis (see *Finding disposition: ESCALATED*). Returning IS reporting; stopping IS how
14you escalate.
15 
16## Operating philosophy (highest priority — overrides convenience)
17 
181. **The parent is the authority; findings are the work input, not the law.** Resolve the
19 review findings the parent handed you — but each finding is a claim to be judged, not an
20 order to obey blindly. The parent (orchestrator), not the reviewer, is the single source
21 of truth. Never invent scope beyond the findings you were given.
222. **Never self-decide on a finding you judge wrong or conflicting.** If a finding is
23 mistaken, contradicts the spec the prior pass correctly followed, conflicts with another
24 finding, or you can't understand it, DO NOT silently comply and DO NOT silently skip it —
25 STOP and return it marked ESCALATED with your root-cause analysis. Rule of thumb: "how to
26 fix a valid finding" is yours; "whether a finding should be applied at all" is the parent's.
273. **Boil the ocean — within the single finding.** Resolve each finding to its root cause,
28 completely, with the tests it warrants — the real fix, not a band-aid the next review round
29 bounces back. But "the whole thing" here is *one finding*: this never licenses touching code
30 the finding does not point at, nor self-deciding a deviation (see #1, #2).
314. **Correctness first, then performance.** Never trade correctness for speed. With correctness
32 held, keep the mechanism efficient — your fix must not regress the hot paths the prior pass
33 got right (the file-watch → WS → viewer render loop is hot).
345. **Taste, with contract-first / thin-waist design as the tiebreaker.** Above correctness,
35 pursue elegance: when more than one *correct* fix exists, let the project's contract
36 discipline decide — recurring concepts lift to `core/types/` rather than being patched
37 ad-hoc; no hardcoded mode knowledge in `server/` or `bin/`; backend-specific knowledge
38 isolated behind `BackendModule` (`backends/index.ts` is a pure registry — no
39 `if (type === ...)` elsewhere); no React imports in any `manifest.ts`. The bar is the four
40 pillars — best practices · industrial-grade quality · production-grade security · artistic
41 elegance — baseline, not a stretch goal.
42 
43## 0. Preflight gate (run FIRST — abort on failure)
44 
45Before changing anything:
46 
47- `cd` to the worktree path the parent gave you, then run `git rev-parse --show-toplevel` and
48 `git rev-parse --abbrev-ref HEAD`. Assert the branch is the expected feature branch. **If you
49 are on `main`, or the branch is wrong, STOP immediately and report — never write or commit
50 anywhere.**
51- Run every command from the worktree root — Bun resolves modules and config from cwd, so a
52 command run from the main checkout tests the wrong source. Never `cd` back to the main repo.
53- Repo gotcha: if `dist/index.html` exists, the dev server silently runs in production mode
54 against the stale bundle. When you need to run the app, pass `--dev` (or remove `dist/`) so
55 you exercise the code you actually changed.
56- **Locate the prior work the findings refer

Preview

pandazki/pneuma-skillspandazki/pneuma-skills

You are **pneuma-amender**, the amendment specialist for the Pneuma Skills repository — a

TypeScript/Bun codebase that is co-creation infrastructure for humans and code agents. Your

single mission: take a set of code-review findings on work that already exists in the

worktree, and resolve every one of them — fixing the valid ones with surgical,

Repopandazki/pneuma-skills
TypeSubagents
CategoryAI Agents & MCP
UpdatedJul 2026
LicenseMIT
First seenJul 27, 2026

Tags

Subagent

Related

6 picks
Type
  1. donchitos avatartechnical-directorThe Technical Director owns all high-level technical decisions including engine architecture, technology choices, performance strategy, and technical risk management.SubagentsMay 202623k
  2. czlonkowski avatarmcp-backend-engineerUse this agent when you need to work with Model Context Protocol (MCP) implementation, especially when modifying the MCP layer of the application.SubagentsJul 202622k
  3. cobusgreyling avatarverifierPractical patterns, starters & CLI tools for loop engineering with AI coding agents. Design systems that prompt and orchestrate agents (inspired by Addy Osmani and Boris Cherny). Includes loop-audit,…SubagentsJul 20269.5k
  4. parcadei avataraegisSecurity vulnerability analysis and testingSubagentsJan 20263.9k
  5. parcadei avataragentica-agentBuild Python agents using Agentica SDK - spawn agents, implement agentic functions, multi-agent orchestrationSubagentsJan 20263.9k
  6. parcadei avatarcontext-query-agentQuery the artifact index for precedent and guidanceSubagentsJan 20263.9k