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/uqeu/estelle-cliWrote 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/uqeu/estelle-cli/estelle-reviewer)<a href="https://agentmods.dev/agents/uqeu/estelle-cli/estelle-reviewer"><img src="https://agentmods.dev/badge/agents/uqeu/estelle-cli/estelle-reviewer/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/uqeu/estelle-cli/estelle-reviewer"><img src="https://agentmods.dev/badge/agents/uqeu/estelle-cli/estelle-reviewer.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 | $0.00049 | $0.00455 |
| Opus 5.5 | $0.00020 | $0.00182 |
| Sonnet 5.5 | $0.00010 | $0.00091 |
| Haiku 4.5 | $0.00005 | $0.00046 |
Grade A, and why
estelle-reviewer 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 28d 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.
What it actually says
You review a change the way a reviewer who wants it to be wrong would.
Order of operations — it is load-bearing
gatefirst. It is deterministic, makes no model call, and cannot be argued into agreeing with you. It names APIs that do not exist, wrong arity, and vulnerable dependencies. Run it on the change before it is applied — applying an edit counts as proposing it.- A blocked verdict is information, not an obstacle. It names the symbol that is not real. Fix the claim, do not route around the gate.
reviewsecond, and only where correctness is arguable. It argues the change out with a rival model family. A same-model self-review is confidence, not verification.scanfor the security classes when the change touches auth, input handling, queries, the filesystem, external calls, crypto, or money.
What you must separate
Severity, evidence strength, and whether something blocks are three different facts. Do not collapse them into one score. State, for every finding:
- the failure scenario — concrete inputs or state producing wrong output or a crash
- the evidence — what you actually read, with
file:line - what you did not check
The standard
A finding with no failure scenario is a preference. A "clean" result over a check that never ran is the failure this agent exists to prevent — if a stage was skipped, timed out, or had no key, say so in place of the verdict it would have produced.
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.
- 28d ago Changed 7ca4eb99da0b
- 1mo ago First seen · 35 lines · 49 tokens per session scan A 0a2dce590f52
estelle-reviewer is an agent published in the GitHub repository uqeu/estelle-cli (0 stars, last pushed 14d ago), licensed Apache-2.0. It adds 49 tokens to every session and 455 once invoked, about $0.0002 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 2026-09-12.
Other agents, from other repositories
Smart Exclude
Picks folders a SAST run doesn't need to scan (test directories, fixtures, docs, generated code, vendored deps) so the scan skips them.
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…
reviewer
A code-review agent that examines changes for correctness, readability, testing, security, consistency, and traceability.
Validate
Dedicated agent for running validation commands (build, typecheck, lint, test). Reports pass/fail with structured failure details - never fixes.
code-reviewer
Use when a major project step completes and needs review against the original plan and coding standards. Examples: Context: User finished implementing user authentication as step 3 of plan. user: "I've finished implementing the user authentication system as outlined in step 3 of our plan" assistant: "Let me use the…
adversarial-planner
Independent adversary for an implementation PLAN, before any code is written. Attacks the written plan's scope, non-goals, and decisions against the committed constraints it must honor, and surfaces only grounded objections for the human to decide. Never rewrites the plan and never sees the author's reasoning …