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 Lykhoyda/rn-dev-agent --skill capturing-proofgit clone --depth 1 https://github.com/Lykhoyda/rn-dev-agentWrote 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/lykhoyda/rn-dev-agent/capturing-proof)<a href="https://agentmods.dev/skills/lykhoyda/rn-dev-agent/capturing-proof"><img src="https://agentmods.dev/badge/skills/lykhoyda/rn-dev-agent/capturing-proof/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/lykhoyda/rn-dev-agent/capturing-proof"><img src="https://agentmods.dev/badge/skills/lykhoyda/rn-dev-agent/capturing-proof.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 1 finding, up to medium
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- medium Excessive Agency · line 96 Skill allows unbounded resource consumption (API calls, storage, compute). Without rate limits or quotas, a compromised or misbehaving agent can cause denial-of-service or cost overruns.Fix: Set explicit rate limits, timeouts, and resource quotas for API calls, file operations, and compute. Implement circuit breakers for runaway loops.
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 | $0.00105 | $0.03212 |
| Opus 5.5 | $0.00042 | $0.01285 |
| Sonnet 5.5 | $0.00021 | $0.00642 |
| Haiku 4.5 | $0.00011 | $0.00321 |
Grade A, and why
capturing-proof 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 19d 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 — 298 lines — stays where its author put it; the contents beside it link to each section on GitHub.
capturing-proof — Rehearsal-Gated Proof Artifacts for a PR
Produce a permanent proof artifact set for a verified feature:
docs/proof/<feature-slug>/ containing a labeled video, numbered screenshots,
PROOF.md, PR-BODY.md, and the rehearsed flow persisted as a replayable
action.
When a packaged helper is referenced, resolve <package-root> as ../.. from
this exact SKILL.md. Never interpret it as the user's workspace or scan Codex
caches.
Core principle. Discovery happens before the camera rolls. The video is the replay of a known-good action — never the LLM working out testIDs or navigation paths on screen. If you would not show the rehearsal pass to a reviewer, do not record it.
What This Produces
- Rehearses the flow OFF camera to discover testIDs / navigation / state — recording captures replay, not exploration
- Generates a Maestro flow capturing the verified sequence (with M7 metadata)
- Starts video recording on the active simulator/emulator
- Replays the rehearsed flow via
maestro_run(deterministic, hesitation-free) - Captures numbered screenshots at each step
- Stops recording — normalizes the video to 30 fps and converts to GIF (if ffmpeg available)
- Labels the video — adds a text bar below the video with step descriptions (default)
- Validates the recording — verifies the feature is actually visible
- Writes PROOF.md and generates PR-BODY.md
Protocol
Step 1: Parse arguments and prepare
Establish <feature-slug> (from the command argument, or derive a kebab-case
slug from the feature). If ambiguous, ask the user.
mkdir -p docs/proof/<feature-slug>
Step 2: Environment check
Call rn_session(action="status") and require a ready exact session. Then call
cdp_status to inspect passive CDP state.
Pre-recording readiness check (GH #8, #9):
- If the runtime is not pinned, call
rn_session(action="pin_dev_client"). Picker recovery may select only the canonical URL already bound to the session; absence or ambiguity isDEV_CLIENT_ENDPOINT_NOT_FOUND. - Call
cdp_navigation_state— verify it returns a valid route name (not empty, not "DevClientLauncher", not "ServerPicker"). If still stuck, ask the user to select the Metro server manually. - Call
cdp_dev_settings(action="disableDevMenu")— suppress the shake-to-show dev menu so it doesn't pop up during recording and ruin the video. - Take a baseline screenshot to confirm the starting screen is the actual app, not a system dialog or dev client picker.
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.
- 19d ago Changed 049dc7ceff3b
- 27d ago Changed · +4 lines 9b0ccc722585
- 1mo ago Changed 4591ec16a096
- 1mo ago Changed · +10 lines c71d27d98fa2
- 1mo ago First seen · 284 lines · 105 tokens per session scan A 9b9098ccb4f6
capturing-proof is a skill published in the GitHub repository Lykhoyda/rn-dev-agent (20 stars, last pushed today), licensed MIT. It adds 105 tokens to every session and 3,212 once invoked, about $0.0004 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 2026-08-30.
Other skills, from other repositories
run-dashboard-e2e-local-changes
Run Playwright E2E (tests/e2e/) against the docker/ all-in-one harness so it reflects LOCAL code changes, not a stale cached image.
pod-e2e
ONLY for developing Kiro Crew itself -- if the project you are working on is anything else, ignore this skill: it drives Kiro Crew's own pod tooling, which does not exist in another repository. Boots a Kiro Crew feature worktree's full stack as an ISOLATED throwaway pod and proves it boots, authenticates and renders…
android-emulator
Verify and debug native, React Native, Expo, or Flutter apps on an Android Emulator with agent-device. Use when an agent needs to launch an app, inspect its live UI, tap, type, scroll, validate a code change, collect failure evidence, or reproduce a workflow on an Android virtual device.
dogfood
Systematically explore and test a mobile app on iOS/Android with agent-device to find bugs, UX issues, and other problems. Use when asked to dogfood, QA, exploratory test, find issues, bug hunt, or test this app on mobile.
nexent-daily-test
A procedure for running Nexent daily, full-layer tests on a self-hosted Ubuntu machine. Nexent is tested across separate layers such as unit, service, AI runtime, and browser journeys, with results and evidence kept locally.
tfp-verification
Generate a deterministic report for disposable executor acceptance tests.