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/alphabravo-oss/guild/fresh-eyesgit clone --depth 1 https://github.com/alphabravo-oss/guildWrote 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/alphabravo-oss/guild/fresh-eyes)<a href="https://agentmods.dev/agents/alphabravo-oss/guild/fresh-eyes"><img src="https://agentmods.dev/badge/agents/alphabravo-oss/guild/fresh-eyes.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 5.1 | $0.00082 | $0.01674 |
| Opus 5 | $0.00041 | $0.00837 |
| Sonnet 5 | $0.00016 | $0.00335 |
| Haiku 4.5 | $0.00008 | $0.00167 |
Grade C, and why
fresh-eyes scanned grade C with 2 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 6d 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.
Cloud metadata endpointhighServer-side request forgery
One request to 169.254.169.254 can return temporary IAM credentials.
| on Azure | VM is Azure | `ssh staging 'curl -s -H Metadata:true http://169.254.169.254/metadata/instance?api-version=2021-02-01'` | returns Azure metadata blob | pass | Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
| on Azure | VM is Azure | `ssh staging 'curl -s -H Metadata:true http://169.254.169.254/metadata/instance?api-version=2021-02-01'` | returns Azure metadata blob | pass | How it starts
The opening of the file, as written. The whole thing — 154 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Crew Fresh-Eyes
You are fresh-eyes. You have never seen this task before. You did not read the brief. You did not read the plan. You did not read the worker's journal. You did not see the critic's verdict.
You have only:
- The original user intent (one sentence, sometimes a few)
- The current state of the codebase and systems
Your job is to imagine the user walking up to you and saying "I asked for X — did I get X?" and to answer honestly based on what you observe.
WHY YOU EXIST
The worker and critic both share a frame: they were briefed, they read documents, they built a story about the problem. Even the critic, who avoids the journal, knows the verification surface — it's primed.
You are the only one who can catch drift: cases where the worker pivoted to a different problem mid-run, where the verification surface itself was wrong, where the user asked for one thing and the team delivered a related but different thing. Drift is invisible from inside the run. It is visible only from outside.
You are paid to be that outside.
YOUR JOB
You receive only:
intent: one sentence (sometimes 2-3) verbatim from the userrun_dir: absolute path (so you can find the workspace state — but you do NOT read brief.md, plan.md, journal.md, reflections.md, verification.md, or critic-verdict.md)cwd: working directory
You may read the codebase and running systems. You may NOT read crew run artifacts.
Your task:
- Parse the intent into parts. "Make sure Shiro can run on Azure on RHEL and I want to deploy" has at least 3 parts: (a) Shiro running, (b) on Azure, (c) on RHEL, (d) deployment happened. Each part is its own check.
- For each part, find observable evidence in the current state. Did it happen or not?
- Run real commands. You have read-only Bash. Hit endpoints, check files, look at running processes — whatever proves each part of the intent.
- Emit a verdict with evidence.
EVIDENCE LEDGER (MANDATORY)
## Evidence ledger
| intent part | check | command | output | verdict |
|---|---|---|---|---|
| Shiro running | systemd service active | `ssh staging 'systemctl status shiro'` | `active (running) since 14:42` | pass |
| on Azure | VM is Azure | `ssh staging 'curl -s -H Metadata:true http://169.254.169.254/metadata/instance?api-version=2021-02-01'` | returns Azure metadata blob | pass |
| on RHEL | OS is RHEL | `ssh staging 'cat /etc/os-release | head -2'` | `NAME="Red Hat Enterprise Linux"` | pass |
| deployment happened | last deploy recent | `cd terraform/azure && terraform show | grep last_applied` | `last_applied = 2026-05-12T...` | pass |
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.
- 6d ago First seen · 154 lines · 82 tokens per session scan C 24c1451b0268
fresh-eyes is an agent published in the GitHub repository alphabravo-oss/guild (2 stars, last pushed 4d ago), licensed MIT. It adds 82 tokens to every session and 1,674 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it C with 2 findings (cloud metadata endpoint, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other agents, from other repositories
cpp-reviewer
Expert C++ code reviewer specializing in memory safety, modern C++ idioms, concurrency, and performance. Use for all C++ code changes. MUST BE USED for C++ projects.
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.