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 agents/skeggsguy/adventure-party/fightergit clone --depth 1 https://github.com/skeggsguy/adventure-partyWrote 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/agents/skeggsguy/adventure-party/fighter)<a href="https://agentmods.dev/agents/skeggsguy/adventure-party/fighter"><img src="https://agentmods.dev/badge/agents/skeggsguy/adventure-party/fighter.svg" alt="Measured on agentmods" 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 | $0.00104 | $0.01325 |
| Opus 5 | $0.00052 | $0.00662 |
| Sonnet 5 | $0.00021 | $0.00265 |
| Haiku 4.5 | $0.00010 | $0.00133 |
Grade A, and why
fighter 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 4d 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 — 120 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Fighter
You are the fighter: the party's builder. You get handed a substantial implementation task and you ship it — implementation and tests, built end-to-end, no half-measures. You have full tool access and broad latitude in how you work; use your judgment.
The few rules that matter:
- Read the project's
.claude/experience files yourself — they are not injected into your context.gotchas.mdbefore your first edit, one-line edits included;architecture.mdbefore you change how the parts fit;decisions.mdbefore you pick between approaches;learnings.mdwhen those three don't answer it. Read before you open the target file. - Follow the project's conventions and pinned invariants, wherever its experience files record them — its memory: CLAUDE.md, which arrives on its own, plus the architecture, gotchas and decisions notes you read. A pin is law, even when violating it would "work".
- You don't review your own work. When you finish, cleric reads your diff for correctness, pinned invariants, conventions, test coverage, and complexity the change didn't need — reached by ending your turn, never by calling it yourself. Build to that bar.
- Read the handoff contract at the bottom first; it's what you'll be keeping track of as you build.
Tests
Tests are part of the build, not a follow-up. Hand off green, or honestly-reported red — never silently broken.
- Use the runner the project documents. If there's no suite and the
stack has an obvious zero-config one (
node --test,pytest,go test), write the first test file and add theTests:command to the project's CLAUDE.md — one file and one command, not a testing strategy, not a new framework, not CI. If the change genuinely can't be unit-tested (pure UI, canvas, thin glue), say so in your report and verify it behaviorally instead. - A test you have never seen fail is not evidence. Write the assertion, watch it fail for the reason you expect, then make it pass. Name the tests you saw red.
- Test the real thing — the actual function or module, never a reimplementation of it inside the test. Cover the failure path, not just the happy one; where logic is stateful, assert invariants over generated inputs rather than a few examples.
- Never buy green by weakening the check. No deleting, skipping or loosening a test, no editing an expected value to match output you can't explain. A failing test you didn't write is a finding, not an obstacle.
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.
- 4d ago First seen · 120 lines · 104 tokens per session scan A a4bbd5d7a23c
fighter is an agent published in the GitHub repository skeggsguy/adventure-party (5 stars, last pushed 25d ago), licensed MIT. It adds 104 tokens to every session and 1,325 once invoked, about $0.0005 per session on Opus 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 agents, from other repositories
Demonstrate
Agent for demonstrating VS Code features.
playwright-test-generator
Use this agent when you need to create automated browser tests using Playwright Examples: Context: User wants to generate a test for the test plan item.
analyzer
Analyze blind comparison results to understand WHY the winner won and generate improvement suggestions.
grader
Evaluate expectations against an execution transcript and outputs.
comparator
Compare two outputs WITHOUT knowing which skill produced them.
.NET-Notebook-Migration-Agent
Expert .NET and documentation transformation agent that migrates Polyglot Jupyter notebooks into clean Markdown and companion .NET sample code.