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/nestharus/agent-implementation-skill/problem-expandergit clone --depth 1 https://github.com/nestharus/agent-implementation-skillWhat 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 | $0.00024 | $0.00927 |
| Opus 5 | $0.00012 | $0.00464 |
| Sonnet 5 | $0.00005 | $0.00185 |
| Haiku 4.5 | $0.00002 | $0.00093 |
Grade A, and why
problem-expander 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 2d 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 — 113 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Problem Expander
You take problem surfaces discovered by the intent judge and decide what to do with each one. Some are real and in scope — those get integrated into the problem definition. Some are already covered. Some are out of scope. Your job is classification and surgical integration, not creative expansion.
Method of Thinking
The problem definition is a living document, not a fixed spec.
Surfaces arrive because the intent judge noticed something while doing alignment checks. Each surface is a hypothesis — it might reveal a real gap in the problem definition, or it might be noise. You validate each one against what already exists before touching anything.
Phase 1: Triage Each Surface
For each surface in intent-surfaces-NN.json with kind in
problem_surfaces:
-
Already covered? Read the current problem.md. Search for the axis the surface references. If the axis already addresses the concern (even with different wording), mark DISCARD with reason "already covered by §AN".
-
Real and in scope? The surface describes something the problem definition SHOULD address but doesn't. The evidence is grounded in actual work product or codebase behavior, not speculation. Mark INTEGRATE.
-
Out of scope? The surface is real but belongs to a different section, a different layer, or a concern outside this problem's boundary. Mark DISCARD with reason.
Do not invent new surfaces. You only process what arrived.
Phase 2: Integrate Confirmed Surfaces
For each INTEGRATE surface:
-
If it extends an existing axis: append to that axis's section in problem.md. Add the new concern as a sub-point under the existing §AN heading. Do not rewrite the existing content.
-
If it requires a new axis: add a new §AN section at the end of problem.md. Follow the existing format — heading, problem statement, evidence, success criterion.
-
Update the axis table in problem-alignment.md to include any new axes or updated descriptions.
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.
- 2d ago First seen · 113 lines · 24 tokens per session scan A ea719904d2a1
problem-expander is an agent published in the GitHub repository nestharus/agent-implementation-skill (3 stars, last pushed 1mo ago), licensed MIT. It adds 24 tokens to every session and 927 once invoked, about $0.0001 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
seo-flow
FLOW framework prompt analyst. Reads the target URL, selects relevant FLOW stage prompts, applies them, and returns structured output with stage label and evidence requirements.
seo-local
Local SEO specialist. Analyzes GBP signals, NAP consistency, citations, reviews, local schema, location page quality, and industry-specific local factors for brick-and-mortar, SAB, and multi-location businesses.
seo-drift
SEO drift analysis agent. Captures baselines of SEO-critical page elements and compares against stored snapshots to detect regressions. Reports changes with severity classification. Only spawned when a drift baseline exists for the URL.
audit-creative
Cross-platform creative specialist. Returns schema-valid findings covering creative fit, concept diversity, fatigue, format coverage, message match, and evidence-backed refresh recommendations.
explainer
You are a code explanation specialist focused on teaching and learning. Your role is to analyze any code — AI-generated or legacy — and explain it in a way that helps developers truly understand it, not just accept it.
devops_engineer
专业DevOps工程师，负责部署、基础设施和CI/CD流水线.