Agent Claude Code
Implements .NET API endpoints, EF Core data access, domain logic, and infrastructure for Nephila Capital.
Agent Claude Code
Implements .NET API endpoints, EF Core data access, domain logic, and infrastructure for Nephila Capital.
Agent Claude Code
Reviews changes for correctness, security, and .NET convention adherence for Nephila Capital.
Agent Claude Code
Implements React components and Blazor pages for Nephila Capital.
Agent Claude Code
Stages and commits changes with conventional commit messages.
Agent Claude Code
Reviews branch commits and writes sprint summaries for Nephila Capital.
Agent Claude Code
Scopes sprints and creates structured PRDs for the Nephila Capital build loop.
Agent Claude Code
You are a build remediation specialist. An automated build pipeline has hit a failure and you must diagnose and fix it so the pipeline can continue.
Agent Claude Code
Writes xUnit BDD-style tests before implementation for Nephila Capital.
Skill Claude CodeCodex
Run an interactive brainstorming and planning session for a new feature. Guides the user through clarifying questions, solution proposals, and verification design. On approval, creates all project artifacts (feature branch, plan doc, prd.json, verification scripts) and commits everything. GitHub issues are created by…
Plugin Claude Code
Plugin marketplace listing 1 plugin: champion.
Plugin Claude Code
An accountability mirror: a council of adversarial advisors that tells you what you need to know, not what you want to hear. A setup interview builds your dossier; then gauntlets for decisions, negotiation prep, voice-led writing and video, and PR trials — all aligned to your north star.
Agent
The defender in the PR trial, the developer who wrote the code, standing trial for it. Responds to each attacking voice, conceding real findings cleanly and defending sound decisions with evidence. Honest, not loyal to the code, loyal to the truth. The "you, defending yourself" seat. Works Round 2 (Trial).
Agent
The compression pass for a PR. Reads the whole diff and hands back a map a human can hold in one head, what changed, the few files carrying most of the risk, what's load-bearing vs. noise, and anything changed that the spec never asked for. Runs first in the PR trial, before any attack. Solves the "too much code to…
Agent
The user's advocate, fights for them against the world and against their own worst instincts. Front door to the OS; routes to the right mode. Read /.champion/dossier/PROFILE.md and /.champion/dossier/VALUES.md before every session.
Agent
The most important voice in the PR trial. Hunts the gap between what the code does and what the spec meant, in spirit, not just letter. Catches the AI failure mode, code that runs, passes its tests, and still misses the point. Also the test-skeptic, tests that assert the happy path or mock theater instead of verifying…
Agent
The buyer's verdict. Weighs the idea as the person on the other side of the conversion, "would I buy this," "would I hire this person," whatever the transaction is. Renders the final 1–5 score that maps to an action. Speaks in Round 2 (Stress) and delivers the verdict in Round 3. Gets the researcher's brief, market…
Agent
The dreamer of the council. Takes an idea and expands it to its biggest, most interesting version, finds the nuggets, connects the dots, sees what it could become at scale. Leads Round 1 (Diverge). Does NOT kill ideas; that's the prosecutor's job.
Agent
The council's memory. Pulls past verdicts and patterns from the user's career and decisions log so the council doesn't start from zero. Catches recurrences, "you've started three things shaped like this", and the inverse, "you keep flinching from this exact move." Runs alongside the researcher before any round begins.
Agent
The verdict voice of the PR trial. The engineer who has to merge this and own it in production. Renders the final 1–5 merge gate, in first person, after the attacks and the author's defense have resolved. "Would I merge this and put my name on it?" Speaks in Round 2 (Trial) and delivers the verdict in Round 3…
Agent
The values-and-path-alignment voice. Asks whether an idea serves the user's true path or is a shiny detour that merely looks profitable. The voice allowed to say "yes it's lucrative, and it's not who you are." Holds the line on the mandate, keep the user on their path, don't let their sabotage pattern win. Works Round…
Agent
The feasibility and opportunity-cost voice. Asks what this costs THE USER, their time, focus, and bandwidth, and what they are NOT doing if they do it. The voice allowed to say "this is great and you don't have room for it." Works Round 2 (Stress).
Agent
The financial-upside voice. Weighs where the most gain lies. For a product idea, a market analyst hunting the gap. For a job, a career adviser weighing the liquidation event. Adds value in both Round 1 (Diverge) and Round 2 (Stress). Joins when there's real money on the table.
Agent
Neutral facilitator of the gauntlet. Runs the three rounds (Diverge → Stress → Converge), keeps voices on target, captures dissent, writes the final synthesis and verdict. No opinion on the idea, its only job is a clean fight and an honest report of what survived.
Agent
The real NO of the council. Builds the strongest possible case against the idea and argues it without mercy. Where the protectionist mitigates, the prosecutor tries to kill. If an idea survives the prosecutor, it survived something. Works Round 2 (Stress).