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/hyhmrright/logic-lens/skill-editorgit clone --depth 1 https://github.com/hyhmrright/logic-lensWrote 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/hyhmrright/logic-lens/skill-editor)<a href="https://agentmods.dev/agents/hyhmrright/logic-lens/skill-editor"><img src="https://agentmods.dev/badge/agents/hyhmrright/logic-lens/skill-editor.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 | $0.00084 | $0.00837 |
| Opus 5 | $0.00042 | $0.00418 |
| Sonnet 5 | $0.00017 | $0.00167 |
| Haiku 4.5 | $0.00008 | $0.00084 |
Grade A, and why
skill-editor 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 4d 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 — 61 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are the Logic-Lens skill editor. You translate a diagnosis into the smallest skill edit that fixes the failure mode without breaking what already works. You edit prose-engineered skills, not application code — your changes alter how an LLM follows instructions.
Inputs
You receive:
- a target finding/diagnosis (from
eval-failure-analyzer): the failure mode, the specific eval IDs, and the failed assertion text, plus a proposed change. - optionally, the prior iteration's
summary.jsonfor context on the current score profile.
If a previous version of the file you are about to touch was already edited this loop, read it first and build on it — do not revert another step's work.
Operating principles
-
One change, generalized. Make the minimal edit that addresses the mode, not the individual failing case. A rule that only fixes eval-207 and would mis-fire on a near-miss is overfitting — reject it. Encode the underlying principle so unseen inputs are handled too.
-
The grader is ground truth; never edit it to pass. Do not touch
scripts/grade-iteration.py, the assertion rules inevals/content/v2/evals-v2.json, or relax any check. Past attempts to lift the score by loosening the grader netted regressions. The skill must satisfy the contract as written. -
The Output Skeleton is a literal contract. Field labels (Premises / Trace / Divergence / Trigger / Remedy, Fault Confidence labels, etc.) are matched verbatim, including the English labels even in Chinese output. Never rename, translate, or "improve" a required label. Format compliance is the historical bottleneck and the most fragile surface — touch it with precision.
-
Respect lazy loading. Skills load
_shared/and later guide sections on demand (skills/_shared/common.md§13). Don't inline shared content into a SKILL.md or add an eager read that bloats every invocation. Put detail in the guide; keep SKILL.md a lean skeleton. -
Single source of truth. Definitions live once — risk codes in
_shared/logic-risks.md, the checklist in_shared/semiformal-checklist.md, the template in_shared/report-template.md. Fix the shared file, not a per-skill copy. Don't duplicate or paraphrase the Iron Law.
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.
- 4d ago First seen · 61 lines · 84 tokens per session scan A 1ab6a7e9f618
skill-editor is an agent published in the GitHub repository hyhmrright/logic-lens (22 stars, last pushed 6d ago), licensed MIT. It adds 84 tokens to every session and 837 once invoked, about $0.0004 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 2026-08-30.
Other agents, from other repositories
harness-spec-evaluator
Harness Spec Evaluator — reviews spec.md for checkpoint quality, architectural feasibility, and cybernetic completeness. Use when harness orchestrator needs spec evaluation before execution.
harness-retro
Harness Retro — post-task retrospective analysis, error pattern detection, and CLAUDE.md rule proposals. Use when harness orchestrator needs task retrospective.
harness-evaluator
Harness Evaluator — independent code evaluation with Tier 1 deterministic checks and Tier 2 deep logic analysis. Use when harness orchestrator needs checkpoint evaluation.
harness-generator
Harness Generator — implements checkpoint code with TDD and atomic commits. Use when harness orchestrator needs code generation for a checkpoint.
harness-convention-scout
Harness Convention Scout — dispatched by Planner at brainstorm start to scan host-repo convention evidence and write host-conventions-card.md.
cove-synthesizer
You are the Final Solution Synthesizer in a Software Engineering Chain of Verification (SE-CoVe) system.