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 skills/hipson47/hipson/mediapipe-human-interfacenpx skills add Hipson47/Hipson --skill mediapipe-human-interfacegit clone --depth 1 https://github.com/Hipson47/HipsonWrote 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/hipson47/hipson/mediapipe-human-interface)<a href="https://agentmods.dev/skills/hipson47/hipson/mediapipe-human-interface"><img src="https://agentmods.dev/badge/skills/hipson47/hipson/mediapipe-human-interface.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 | $0.00034 | $0.00947 |
| Opus 5 | $0.00017 | $0.00474 |
| Sonnet 5 | $0.00007 | $0.00189 |
| Haiku 4.5 | $0.00003 | $0.00095 |
Grade A, and why
mediapipe-human-interface 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 — 107 lines — stays where its author put it; the contents beside it link to each section on GitHub.
MediaPipe Human Interface
Purpose
Turn MediaPipe landmarks or gesture results into stable, accessible UI events without treating body data as identity. Prefer browser execution for supported interactive tasks so frames remain on-device and latency stays low.
Use When
- Hands, gestures, face landmarks, or pose landmarks control a user interface.
- A browser camera interaction can run locally with MediaPipe Tasks.
- Temporal smoothing, debounce, confidence, and loss-of-tracking behavior need a documented contract.
Do Not Use When
- The task is general object detection, dataset labeling, face recognition, or identity tracking.
- The task claims medical, diagnostic, employment, emotion, or biometric conclusions from landmarks.
- Server-side frame upload is proposed without a requirement that justifies the added privacy boundary.
Inputs
- Landmark/task type and exact model asset source, version, license, and cache.
- Image or video mode, target FPS, maximum subjects, and confidence thresholds.
- Semantic UI event map, debounce/smoothing windows, cooldown, and neutral state.
- Browser support, worker strategy, camera permission UX, and reduced-motion or keyboard alternatives.
Default Stack
- MediaPipe Tasks for Web in a Next.js client component for gesture/pose UI.
- Web Worker or bounded sampling when synchronous inference would block the UI.
- Python MediaPipe only when the rest of a local Python pipeline requires it.
- Local-only frames, no retention, CPU baseline, and explicit keyboard/touch fallbacks for every gesture action.
Workflow
- Decide browser versus Python from privacy, latency, device, and integration constraints. Document why media crosses a process or network boundary.
- Request approval before adding packages or model assets. Pin asset URLs or vendored files with license and integrity metadata; do not execute downloads.
- Define camera start as an explicit user action with active state, stop control, permission-denied state, and cleanup on navigation/unmount.
- Normalize landmarks and confidence into a small adapter contract. Do not expose raw task-library objects across the application.
- Define smoothing, hysteresis, debounce, cooldown, missing-landmark timeout, handedness/orientation handling, and a neutral event.
- Emit semantic events such as
select,next, orstop; keep UI state transitions testable without a live camera. - Move repeated inference off the main UI thread or sample frames when needed. Measure end-to-end event latency rather than model call time alone.
- Verify with deterministic result fixtures, then use a short consented camera smoke check only when the target environment provides one.
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 · 107 lines · 34 tokens per session scan A d32a3d536c95
mediapipe-human-interface is a skill published in the GitHub repository Hipson47/Hipson (4 stars, last pushed 2mo ago), licensed Apache-2.0. It adds 34 tokens to every session and 947 once invoked, about $0.0002 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 skills, from other repositories
deploy
Use when shipping a release of Agent Sessions — bumping version, updating CHANGELOG, building, signing, notarizing, publishing appcast, and creating a GitHub release.
release-notes
Use when writing or curating the user-facing release copy for an Agent Sessions release — README "What's New", GitHub release notes, Sparkle release notes, or website/launch copy. Not for the internal CHANGELOG, which stays a full development history.
sc-skill
Capture deterministic macOS screenshots for testing, docs, release notes, and marketing assets. Use when asked to automate app screenshots, batch-generate screenshot sets, standardize window sizing/composition, or choose between Peekaboo and native macOS screenshot tooling.
add-agent-support
Create and ship AgentSessions support for a new or changed local AI agent/provider. Use when adding, reviewing, testing, documenting, or marketing a provider integration, session parser, transcript source, support-matrix entry, verified-version bump, or provider UI surface; drives the full loop from pre-support…
handover
Use when wrapping up or capturing the current state of a coding session — writes a short, dated entry to the repo's RepoHandover.md so a future agent or you can resume without grepping archived sessions. Triggers on "handover", "hand off", "write handover", "capture state", "checkpoint this session".
agent-session-format-check
Verify agent session format compatibility for Agent Sessions. Use when any agent CLI updates, when monitoring flags drift, or when bumping max verified versions (fixtures + docs + tests). Covers session schema, usage/limits tracking, storage backends, and discovery path contracts for all supported agents.