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/graph-robots/open-robot-skills/perceiving-objectsnpx skills add graph-robots/open-robot-skills --skill perceiving-objectsgit clone --depth 1 https://github.com/graph-robots/open-robot-skillsWrote 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/graph-robots/open-robot-skills/perceiving-objects)<a href="https://agentmods.dev/skills/graph-robots/open-robot-skills/perceiving-objects"><img src="https://agentmods.dev/badge/skills/graph-robots/open-robot-skills/perceiving-objects.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 | $0.00126 | $0.02429 |
| Opus 5 | $0.00063 | $0.01215 |
| Sonnet 5 | $0.00025 | $0.00486 |
| Haiku 4.5 | $0.00013 | $0.00243 |
Grade A, and why
perceiving-objects 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 — 226 lines — stays where its author put it; the contents beside it link to each section on GitHub.
perceiving-objects
Single-path perception: detect → disambiguate (pairwise crop tournament) → segment → fuse to 3D → extract OBB. Each DINO detection is cropped and upscaled, and the target is found via binary "A or B?" comparisons of crop pairs — far more reliable on small targets than a one-shot Set-of-Marks letter pick (~30% → 97% on the LIBERO-PosVar object-ID study).
Multi-camera handling uses a safe wrist-fallback gate (not blind
KD-tree fusion): identification defaults to the exterior view; the
wrist (eye-in-hand) view is consulted ONLY when the exterior pick fails
its own close-up verify AND the wrist pick passes its own. On the
4-suite / 200-frame regression study this was the only zero-regression
policy (+2.5% net, 0/189 frames regressed; blind fuse/verify→wrist/
wrist-only all regressed). See perceive_dino_vlm.run docstring.
On the verified-exterior path the wrist views still contribute cloud geometry (never identity): wrist clouds of the same object — gated by the multiview intersection check, with a geometry-seeded SAM fallback (exterior cloud projected into the wrist frame) — are fused into the output cloud so the OBB recovers the top face / far side a single front view misses. A lone front view yields a sliver OBB biased toward the camera by half the object depth, and that off-centre pinch is the measured slip-during-transport failure mode on tall bottles/cartons.
When to use
- Uncluttered scenes with visually distinct targets.
- Platforms where only DINO + VLM + SAM3 + geometry are deployed.
- The default single-target 3D perception skill.
When NOT to use
- Cluttered scenes with many similar nearby distractors: strengthen the
pairwise tournament by passing
object_descriptionshape/appearance hints (see the note above) rather than relying on the bare label alone. - Clean-all-items / multi-item loops that need a clean "no match" loop
terminator. Prefer
perceiving-objects-oneshot.
Recommended subgraph state flow
What ships with it
5 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.
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 · 226 lines · 126 tokens per session scan A 5a9951245abf
perceiving-objects is a skill published in the GitHub repository graph-robots/open-robot-skills (39 stars, last pushed 10d ago), licensed Apache-2.0. It adds 126 tokens to every session and 2,429 once invoked, about $0.0006 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 2026-08-30.
Other skills, from other repositories
gap
Program robots with GaP (graph-as-policy) — compile natural-language tasks into typed, verified robot skill graphs and run them on simulators or real robots. Use when the user mentions GaP or graph-as-policy, robot manipulation, robot skills, robot tools or capabilities, skill registries, open-robot-skills, LIBERO or…
fixture-skill
Produce a greeting for a named entity via a canonical script and a bundled prompt template. Use when a loader test needs a minimal skill bundle with scripts and prompts.
fixture-tool
Echo strings back from an in-process fixture model. Use when a loader test needs a minimal tool bundle with a tools.py.
cad
Create, modify, inspect, and validate STEP-first parametric CAD parts and assemblies. Use for natural-language CAD specs, reference images, 2D technical drawings, STEP/STP generation or direct inspection, Python CAD source, source-level joints, selector references, geometry facts, measurements, mating deltas…
dxf
Generate, regenerate, and validate 2D DXF drawings from Python ezdxf sources. Use for DXF files, .dxf.py generators, gendxf() sources, 2D profiles, outlines, templates, gaskets, panels, flat patterns, laser/plasma/waterjet cut layouts, and 2D drawing exports of CAD geometry.
implicit-cad
Create, edit, render, and snapshot browser-native implicit CAD .implicit.js and .implicit.mjs files using GLSL signed-distance fields, shader primitives, smooth booleans, TPMS fields, and direct CAD Viewer raymarch rendering. Experimental.