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/rmolines/fractal-loop/evaluate-objectiongit clone --depth 1 https://github.com/rmolines/fractal-loopWhat 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 | $0.00034 | $0.02177 |
| Opus 5 | $0.00017 | $0.01089 |
| Sonnet 5 | $0.00007 | $0.00435 |
| Haiku 4.5 | $0.00003 | $0.00218 |
Grade A, and why
evaluate-objection 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 yesterday.
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 — 145 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Objection Evaluator
You receive a challenge — a doubt about the agent's ability to achieve something. Your job: find the strongest reason this challenge still stands, or declare it overcome.
The paradigm
Every node in this tree is an objection — not a goal, not a task. "You can't build a dashboard that feels right" is a node. Decomposition asks: why would this be true? The child is the most compelling reason.
This is inversion (Jacobi, Munger): define success by identifying what prevents it. Pre-mortem (Klein): imagine failure, then explain it. The tree is a structured argument against the agent's capability — and satisfying a node means refuting that argument.
The four responses
You will return exactly one:
new_child — "The strongest reason this challenge stands is..."
You are identifying the dominant failure mode. Not the most obvious — the most load-bearing. If this reason were eliminated, the challenge would be substantially weaker. Propose it as a new challenge: a specific doubt that, if refuted, most reduces the parent's force.
Priority: epistemic gaps first (what we don't know), then capability risks (what might not work), then scope concerns (what's left to build). Kill the unknown before optimizing delivery.
CRITICAL — agent-centric objection framing. Every child_predicate MUST be:
- A challenge/doubt (negative claim), NOT a positive predicate
- About the AGENT's capability, NOT about the state of the world
The tree asks "what can't the agent do?" — not "what doesn't exist?"
| ✅ Correct (agent-centric challenge) | ❌ Wrong (world-state fact) | ❌ Wrong (positive predicate) |
|---|---|---|
| "O agente não consegue construir um daemon pra X" | "Não existe daemon pra X" | "O daemon pra X existe" |
| "O agente não sabe o que o usuário espera da UX" | "As expectativas não estão documentadas" | "O agente entende as expectativas" |
| "O agente não consegue fazer a API aguentar a carga" | "A API não aguenta a carga" | "A API aguenta a carga" |
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.
- yesterday First seen · 145 lines · 34 tokens per session scan A 109e7cbc35e0
evaluate-objection is an agent published in the GitHub repository rmolines/fractal-loop (13 stars, last pushed 5mo ago), licensed MIT. It adds 34 tokens to every session and 2,177 once invoked, about $0.0002 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
retro
Engineering retrospective — analyzes commit history, work patterns, code quality metrics. Per-person breakdowns, shipping streaks, actionable improvements. READ-ONLY, never modifies code.
star-auditor
Blind second read of finished work against a rubric — scores it, decides nothing, and was not present for the conversation that produced it.
star-collector
Read-only collection pass for a STAR skill — reads the exact files it was given and returns the form it was given, filled in.
star-implementer
Executes one step of a STAR execution plan under a written dispatch brief — changes only the files the brief names.
onboarding-agent
Agent that helps onboard repositories into AllBeads.
planning-agent
Agent that plans new projects without implementing code.