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.
git clone --depth 1 https://github.com/JackScammell/Parliament-Of-ChaosWrote 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/commands/jackscammell/parliament-of-chaos/debate-replay)<a href="https://agentmods.dev/commands/jackscammell/parliament-of-chaos/debate-replay"><img src="https://agentmods.dev/badge/commands/jackscammell/parliament-of-chaos/debate-replay/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/commands/jackscammell/parliament-of-chaos/debate-replay"><img src="https://agentmods.dev/badge/commands/jackscammell/parliament-of-chaos/debate-replay.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 | $0.00018 | $0.01134 |
| Opus 5.5 | $0.00007 | $0.00454 |
| Sonnet 5.5 | $0.00004 | $0.00227 |
| Haiku 4.5 | $0.00002 | $0.00113 |
Grade A, and why
debate-replay scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 97 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Debate Replay
Replay a past /debate-topic session from its snapshot and verify that the deliberation engine produces the same outcome. If the replay diverges, either the engine regressed or a model change has perturbed convergence — both are actionable signals.
This command depends on /session-snapshot (Tier 4 — item 4.1). Without a snapshot, there is nothing to replay; the command will report clearly when snapshot infrastructure is absent.
Usage
/debate-replay <session-id> [--snapshot <path>] [--strict] [--compare-to <session-id>]
Examples:
/debate-replay fc777472
/debate-replay fc777472 --strict # Fail loudly on any divergence
/debate-replay fc777472 --compare-to ea113388 # Compare two runs of the same topic
Options
<session-id>(required): The original debate session to replay.--snapshot <path>(optional): Explicit path to the snapshot. Defaults to${CLAUDE_PLUGIN_DATA}/sessions/<session-id>/snapshot.json.--strict(optional): Exit non-zero if the replay's verdict, convergence trajectory, or vote distribution differs from the original.--compare-to <session-id>(optional): Compare against a different session rather than the snapshot's own original. Useful for A/B testing prompt changes.
Process
- Load snapshot
- Resolve the snapshot path, default
${CLAUDE_PLUGIN_DATA}/sessions/<id>/snapshot.json - Fail clearly with a pointer to
/session-snapshotif missing
- Resolve the snapshot path, default
- Recreate context
- Read the snapshot's
topic,mode,participants,seed, andround_inputs - Restore the agent roster exactly — same versions, same effort tiers, same
disallowedTools
- Read the snapshot's
- Run the debate
- Invoke
deliberation-conductorwith the recorded topic and mode - Feed the same seed to the sampler for determinism (model sampling is inherently non-deterministic, but deterministic seeds reduce variance)
- Invoke
- Diff the outcomes
- Compare per-round: novelty, overlap, convergence metrics
- Compare final votes and verdict
- Compute a divergence score: fraction of per-agent positions that shifted verdict
- Report
- Side-by-side table of original vs replay for each round
- Highlight divergence points and classify (stochastic noise vs material disagreement)
- Persist
- Write the replay transcript to
${CLAUDE_PLUGIN_DATA}/sessions/<id>/replays/YYYY-MM-DDTHH-MM-SSZ.json - Update
${CLAUDE_PLUGIN_DATA}/sessions/<id>/replays/INDEX.jsonfor future comparisons
- Write the replay transcript to
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.
- 1mo ago First seen · 97 lines · 18 tokens per session scan A 216ad681e4c9
debate-replay is a command published in the GitHub repository JackScammell/Parliament-Of-Chaos (4 stars, last pushed yesterday), licensed MIT. It adds 18 tokens to every session and 1,134 once invoked, about $0.0001 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 2026-08-31.
Other commands, from other repositories
session
Complete guide to managing pair programming sessions.
expect
Diff-aware AI browser testing — reads the git diff, maps changes to affected pages via the route map, generates a targeted test plan, and executes it via agent-browser (Rust daemon + CDP, ARIA-tree-first) with pass/fail reporting. Use when testing UI changes, verifying PRs before merge, or running regression checks on…
generate-widget
Génération Widget Flutter avec Tests.
test-suite
Run comprehensive test suite with coverage analysis.
email-test-suite
../../services/email/email-test-suite.md.
AGENTS
Run the current binary in a temporary test repository with a real PTY. Use tmux-cli when available, capture terminal output, and inspect -vv logs when behavior depends on timing or shell state. Avoid treating a piped command as proof of interactive behavior.