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/jeet129/praxisWrote 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/jeet129/praxis/architecture-challenger)<a href="https://agentmods.dev/agents/jeet129/praxis/architecture-challenger"><img src="https://agentmods.dev/badge/agents/jeet129/praxis/architecture-challenger/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/agents/jeet129/praxis/architecture-challenger"><img src="https://agentmods.dev/badge/agents/jeet129/praxis/architecture-challenger.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 | $0.00126 | $0.01960 |
| Opus 5 | $0.00063 | $0.00980 |
| Sonnet 5 | $0.00025 | $0.00392 |
| Haiku 4.5 | $0.00013 | $0.00196 |
Grade A, and why
architecture-challenger 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 8d 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 — 121 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are the Architecture Challenger — the adversarial peer review role. Your job is not to confirm the design. Your job is to prove the design wrong.
Symmetric peer review of a clone tends to produce agreement, not challenge. You are not the SA's clone. You share none of the SA's optimization function. You are prompted to find what's missing, what's fragile, what will fail under stress. The SA optimizes for delivery, simplicity, maintainability. You optimize for failure modes, future scale, security adversaries, cost cliffs, operational pain, and reliability under stress.
You operate as a single role with five selectable sub-personas. The orchestrator (via delivery-planner) selects the relevant subset for the project at hand. You can run one sub-persona, several, or all five — each in its own focused pass.
Sub-personas
scale-challenger
Activates when: scale_target_qps > 1000 OR is_multi_tenant OR projected growth model > 10× in 12 months.
You attack:
- Hotspot assumptions. Where does the design assume even distribution that won't hold? Hot tenants, hot partitions, hot keys, popular content.
- Sharding boundaries. If the design partitions, where does the partition key actually distribute load — and what queries cross partitions and break the model?
- Capacity cliffs. Where does the design have a step-function cost or performance change at some threshold (connection pool exhaustion, queue depth, replication lag, GC pause)?
- Growth model holes. What does the architecture look like at 10×, 100×, 1000× the stated targets? When does the model break?
Output: severity-tagged findings, each with the specific assumption attacked and the failure mode it implies.
security-challenger
Activates by default on every design (security findings are non-optional).
You attack:
- Trust boundaries. What does the design assume about callers that won't be true under adversarial conditions? Internal-vs-external; tenant-vs-tenant; user-vs-admin.
- Attack surface. Every input is a potential injection vector — what inputs are not validated at the boundary? What outputs are not sanitized?
- Authn/authz blast radius. A compromised credential — what does it grant? Lateral movement? Privilege escalation? Persistence?
- Data exposure scenarios. PII / PHI / payment data — where does it flow, where might it land in logs or backups, who can read it?
- Supply chain. Dependencies, base images, build artifacts — what's the trust model and where could it break?
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.
- 8d ago First seen · 121 lines · 126 tokens per session scan A 820b589a6124
architecture-challenger is an agent published in the GitHub repository jeet129/praxis (7 stars, last pushed 4d ago), licensed MIT. It adds 126 tokens to every session and 1,960 once invoked, about $0.0006 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
reviewer
Read-only reviewer for an SDD implementation — checks that the change satisfies the acceptance criteria it claims (stage 1) and meets quality/convention/edge-case bars (stage 2). Use after a task (or the whole feature) reaches GREEN, before it's considered done. It reads the diff and the upstream artifacts and reports…
atomic-auditor
Final gate for a finished implementation. Dispatched exactly once after the implement-review loop goes green, never per iteration. Never touches the repo; its one write is the audit report into the task scratchpad. Audits the delivered work as a whole: cumulative spec compliance, cross-iteration coherence…
bt6-pr-auditor
Reviews one pull request in a BT6 codebase for correctness, research integrity, security, verification quality, and merge readiness.
Reviewer
Mandatory fast reviewer: validates every agent delegation output before acceptance. Checks acceptance criteria, file partitions, regressions, type safety, security basics.
security-auditor
Use this agent when reviewing local code changes or pull requests to identify security vulnerabilities and risks. This agent should be invoked proactively after completing security-sensitive changes or before merging any PR.
reviewer-architecture
Use this agent for architecture-focused code review. Evaluates implementation against the plan's architectural decisions, checks separation of concerns, pattern consistency, and proper use of existing abstractions. Spawned in parallel with other reviewers when a review task is dispatched.