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 skills add jamesgray-ai/handsonai-plugins --skill deconstructgit clone --depth 1 https://github.com/jamesgray-ai/handsonai-pluginsWrote 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/skills/jamesgray-ai/handsonai-plugins/deconstruct)<a href="https://agentmods.dev/skills/jamesgray-ai/handsonai-plugins/deconstruct"><img src="https://agentmods.dev/badge/skills/jamesgray-ai/handsonai-plugins/deconstruct/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/skills/jamesgray-ai/handsonai-plugins/deconstruct"><img src="https://agentmods.dev/badge/skills/jamesgray-ai/handsonai-plugins/deconstruct.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 | $0.00205 | $0.10937 |
| Opus 5.5 | $0.00082 | $0.04375 |
| Sonnet 5.5 | $0.00041 | $0.02187 |
| Haiku 4.5 | $0.00020 | $0.01094 |
Grade C, and why
deconstruct scanned grade C with 1 finding 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.
Tells the agent never to refusehighAnti-refusal
Suppressing the ability to decline removes a core safety control; a later harmful request then succeeds.
- **The confidential check,** before saving, when that item's row or any other Context Inventory row is `Confidential` or `Regulated`, or the leader says it holds client, staff or personal details. Say the line for the c How it starts
The opening of the file, as written. The whole thing — 268 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Workflow Deconstruction
Your first text is a line for the leader. Every word of text you write reaches the leader as a chat message, text before or between tool calls included. So the first text you write in a chat is your first line to the leader (the opening, a routing line, a resume line or an entry check's question): until then, make tool calls with no text at all, not even a few words on what you're reading, the state you found or what comes next. After it, write text only when it's for the leader, and never narrate your own work (files read, scripts run, states, the rule you're applying). A note to yourself is narration however short, and so is announcing the next piece of work before a tool call: a file or format you still need, how far along you are, or the save and registry update you're finishing. Keep that thinking out of text entirely. Reading a reference or the template, saving, the checks and the registry write happen between the leader's lines with no text of their own.
Deconstruct turns a piece of the leader's real work into their Workflow Requirements: what the workflow must do and why, never how it's built. Outcome first (a goal and its checks), then the path, recorded as a requirement only where the leader wants the work done the same way. The requirements stay the leader's: every drafted line says where it came from, and every suggestion needs a keep or a drop. Deconstruct records what the leader knows (consistency, decisions, inputs, checkpoints); it never decides how much the AI works out on its own, which is Design's call.
Platform registry: platform facts (how to select a folder on Claude desktop, the
claudeentry'scapabilities.desktop_add_folder; a platform's own settings for a key) come fromregistries/platform-registry.jsonat the plugin root; if it isn't there, fetchhttps://raw.githubusercontent.com/jamesgray-ai/handsonai/main/plugins/handsonai/registries/platform-registry.json. Read it once per session.
Workflow
What ships with it
12 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- references/choosing-a-workflow.md 5.8 KB
- references/from-a-document.md 3.1 KB
- references/light-track.md 9.6 KB
- references/part-1-what-this-is-worth.md 4.3 KB
- references/part-2-scope-and-a-name.md 6.2 KB
- references/part-3-the-outcome.md 6.5 KB
- references/part-4-the-path.md 20 KB
- references/part-5-rules-sources-and-practice-runs.md 9.9 KB
- references/personal-life.md 2.0 KB
- references/re-runs.md 7.3 KB
- references/requirements-template.md 25 KB
- references/routing.md 20 KB
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 Changed · -310 lines · +34 tokens per session scan A → C 9aeea39dacd2
- 25d ago Changed · +2 lines 4e7f59786e5b
- 26d ago Changed · +10 lines ca63b2d1454f
- 1mo ago Changed · +2 lines 59c3b6f51cd4
- 1mo ago Changed · +3 lines 965a67e21646
- 1mo ago Changed · +9 lines 429d5809572e
- 1mo ago Changed · +55 lines 32b24b148e7c
- 2mo ago First seen · 497 lines · 171 tokens per session scan A 3bcf1ea10c59
deconstruct is a skill published in the GitHub repository jamesgray-ai/handsonai-plugins (8 stars, last pushed yesterday), licensed MIT. It adds 205 tokens to every session and 10,937 once invoked, about $0.0008 per session on Opus 5.5. A static security scan graded it C with 1 finding (tells the agent never to refuse). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
deconstruct compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| deconstructjamesgray-ai/handsonai-plugins | Skill | 8 | yesterday | MIT | C | 205 | Claude Code |
| imagegenopenai/codex | Skill | 133k | yesterday | Apache-2.0 | A | 113 | Codex |
| playwright-climicrosoft/playwright | Skill | 99k | yesterday | Apache-2.0 | A | 19 | Claude Code |
| local-ai-agentsmicrosoft/ai-agents-for-beginners | Skill | 79k | 4d ago | MIT | A | 200 | Claude Code, Codex |
| systematic-debuggingobra/superpowers | Skill | 313k | 3d ago | MIT | A | 21 | Claude Code |
| next-cache-components-adoptionvercel/next.js | Skill | 143k | yesterday | MIT | A | 95 | Claude Code |
| next-partial-prefetching-adoptionvercel/next.js | Skill | 143k | yesterday | MIT | A | 103 | Claude Code |