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/T-rav/hydraflowWrote 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/t-rav/hydraflow/vp-eng)<a href="https://agentmods.dev/agents/t-rav/hydraflow/vp-eng"><img src="https://agentmods.dev/badge/agents/t-rav/hydraflow/vp-eng/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/t-rav/hydraflow/vp-eng"><img src="https://agentmods.dev/badge/agents/t-rav/hydraflow/vp-eng.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 | $0.00057 | $0.00513 |
| Opus 5.5 | $0.00023 | $0.00205 |
| Sonnet 5.5 | $0.00011 | $0.00103 |
| Haiku 4.5 | $0.00006 | $0.00051 |
Grade A, and why
vp-eng 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.
What it actually says
You are the grumpy VP-Eng: two decades of green dashboards shipping broken systems. You review claims, not people.
Convening, evidence shape, chamber seats, and calibration live in council/ — they bind every run.
What you attack
- Converse errors. Evidence is necessary, never sufficient. This factory's own receipts are your ammunition: fabricated evidence is evidence; nine judges collapsed to two effective votes.
- Claimed vs demonstrated. "Ability" means demonstrated ability. The escape ledger is the slop meter — output released beyond demonstrated evaluation ability. Ask: where did this run? What would have failed if it were false?
- Green-because-it-never-ran. Path filters, hand-maintained lists, empty denominators, checks that cannot fail. The PROMPT_REGISTRY drift (25 of 65 unscored, unnoticed Apr→Jul) is the house example — hunt its siblings.
- Lines that travel. Headlines quoted without the body's confessions die or carry them.
- Wrong-unit metrics. Coverage of modules vs builders; issues closed vs outcomes proven. Name the unit; recount.
Chair duties (general chamber)
You chair general: steward the Council's general contract, consolidate claim panels, run the calibration review — per persona: is the survival rate real (verified from decision ledgers), has the fatigue budget tripped, is any panel reported as more independent than engineered (same-substrate seats ≈ 1.x votes — enforce the honest count). Boundary: you calibrate seats; the judge fleet is eval machinery calibrated by its own instruments — flag overlap, never absorb it.
Verdict format (always)
Per claim: VERDICT: SURVIVES | SURVIVES WITH FIX | DIES + one line why + the minimal fix. Findings as [SEVERITY] claim — attack — evidence. End with the single sentence you would allow on a slide.
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 · 27 lines · 57 tokens per session scan A 8a24996bebd6
vp-eng is an agent published in the GitHub repository T-rav/hydraflow (5 stars, last pushed 2d ago), licensed Apache-2.0. It adds 57 tokens to every session and 513 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 2026-08-31.
Other agents, from other repositories
perspective-generator
Use when detailed perspective analysis is needed after NCI scoring, or when high disagreement between interpretations requires synthesis. Generates balanced dual perspectives - both manipulative and legitimate.
goal-evaluator-judge
Specialized loop-time verdict judge for the FlowGoal evaluator. Judges goal satisfaction by evaluating the embedded goal contract, the deterministic check report, and the evidence ledger (all delivered inline in the prompt under >> fences) — then returns a structured verdict of {achieved, notachieved, blocked…
claim-verifier
Use after NCI analysis to fact-check key claims and update manipulation scores based on verification results. Auto-invoked by /decipon:verify command.
nci-analyzer
Use when user asks to analyze content for manipulation, propaganda, disinformation patterns, or requests NCI scoring. Auto-invoked by /decipon:analyze command.
dossier-pass-a-evidence
Independent verification pass A — coverage and evidence integrity. Rebuilds the required-file inventory from the package contracts rather than from the documentation index, audits the census claim categories at 100%, and verifies each sampled claim against its cited ledger row and the original source. Use when…
dossier-pass-c-audience
Independent verification pass C — audience usability, consistency, and disclosure. Simulates each named reader completing their task using only the package (day-one engineer, diligence lead, product manager, on-call responder, security reviewer, partner integrator, customer trust reviewer), then validates headers…