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 iurysza/android-use --skill readbackgit clone --depth 1 https://github.com/iurysza/android-useWrote 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/iurysza/android-use/readback)<a href="https://agentmods.dev/skills/iurysza/android-use/readback"><img src="https://agentmods.dev/badge/skills/iurysza/android-use/readback/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/iurysza/android-use/readback"><img src="https://agentmods.dev/badge/skills/iurysza/android-use/readback.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 | $0.00012 | $0.00397 |
| Opus 5.5 | $0.00005 | $0.00159 |
| Sonnet 5.5 | $0.00002 | $0.00079 |
| Haiku 4.5 | $0.00001 | $0.00040 |
Grade A, and why
readback 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
Definition
A readback repeats an instruction, request, or piece of information back to the user. It states what the agent believes the user asked or meant, so the user can confirm or correct the understanding before work continues.
Use it to catch misunderstandings before they become wrong work.
Purpose
- Confirm that the agent correctly received and understood the user's instruction or intent.
- Allow the user to verify accuracy and correct any misunderstanding.
- Prevent errors caused by acting on a mistaken interpretation.
Process
- Inspect the relevant context.
- Infer the user's intention or meaning.
- Keep important constraints and conditions.
- Name assumptions or unresolved ambiguity.
- Give the readback.
- Stop and wait for confirmation or correction.
Domain Language + Plain-English Gloss
Use established terms of art when they accurately describe the user's intent, but immediately gloss them in plain English.
Prefer:
Use a vertical slice — build one small part of the feature end-to-end so it can actually be tested.
Over either:
Use a vertical slice.
or:
Build one small part through every layer of the system...
The idea is to preserve useful domain language without requiring the user to already know it.
Do not introduce jargon just to sound technical. Use it when it gives the idea a useful name.
Rules
- Do not turn the readback into an implementation plan.
- Do not silently resolve meaningful ambiguity.
- Preserve conditions closely enough that their meaning cannot be lost.
- Use domain language when useful, followed by a plain-English gloss.
- If the user corrects the readback, update it before proceeding.
Principle
Make sure your response sounds like a person talking to another. Drop jargon, ceremony, repetition, and unnecessary structure. Preserve the meaning and try to match user's tone/voice.
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 · 61 lines · 12 tokens per session scan A fe3c3311ebf3
readback is a skill published in the GitHub repository iurysza/android-use (21 stars, last pushed 16d ago), licensed MIT. It adds 12 tokens to every session and 397 once invoked, about $0.0000 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 2026-09-15.
Other skills, from other repositories
babysit
Use when a user asks to babysit, monitor, keep checking, keep an eye on, or report when a pull request, CI run, ticket, deployment, or other changing target reaches an outcome.
rubber-duck
Adversarial "rubber duck" review that turns explaining-out-loud into a hallucination check. The main session is the PRESENTER (it did the work — a design doc, investigation, or analysis — and holds the real reasoning) and reconstructs the topic to a LISTENER — a spawned subagent pinned to a DIFFERENT-vendor model that…
building-agents
Use when building or restructuring an LLM agent — provider adapter, tool calling, structured output, RAG, agent loop, eval gate, cost routing, tracing, MCP server — model-agnostic across OpenAI/Anthropic/Gemini/OSS so a model swap is a config change. NOT vector-store SQL alone (that is postgresdb) or service…
agent-eval
Use when measuring whether an LLM or agent system actually got better and gating merges on it: golden sets, fixing an inflated LLM-as-judge, scoring RAG (faithfulness, contextual recall) or agent trajectories (tool correctness, completion), or picking an eval framework. NOT building the agent loop, tools or RAG…
browser-qa
Testing web apps through a real browser: reconnaissance before acting, selector discovery, state capture, console errors and screenshot evidence. Use on a web surface end to end, or to reproduce a front-end bug.
code-review
Reviewing code for defects: diff-first reading, severity classification, evidence over inference, finding caps, and terminal verdicts. Use when reviewing a diff, PR, or patch, or auditing someone else's work.