CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

perception-localize

Locate objects with a running PerceptionLocalize stack over MCP (`state`, `localize`) - probe the machine, build the time window, watch labels on a schedule, narrow an object's position, change params per call, and tell where a failure comes from. Use whenever a blueprint with PerceptionLocalize and McpServer runs, such as go2-localize or xarm-localize-live.

Quality

92%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Passed

No findings from the security scan

SKILL.md
Quality
Evals
Security

Quality

Content

88%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

An excellent operational reference: concrete executable commands, realistic output examples, explicit preconditions, validation checkpoints, and error-recovery guidance throughout. The only refinements available are deduplicating two repeated facts and moving the params/failure reference tables into a separate bundle file.

Suggestions

State the laptop RTX 5070 8 GB reference benchmark once (in "Labels and GPU") and reference it from "Watching labels on a schedule" instead of repeating it.

Move the LocalizeParams field table and the per-rig failure diagnosis into a references/ file, keeping SKILL.md as the operational overview with clearly signaled pointers.

DimensionReasoningScore

Conciseness

The body is dense and stack-specific with no padding about concepts Claude already knows; nearly every line carries non-obvious operational facts (frame cadence, timeout behavior, caching rules). It is not a 5 because of small duplications: the laptop RTX 5070 8 GB benchmark appears in both "Watching labels on a schedule" and "Labels and GPU", and the fuse_voxel_m desk values are stated in two places.

4 / 5

Actionability

Every section is built around copy-paste-ready commands (`uv run dimos mcp call localize -j '...' --timeout 300`), real annotated output samples, and tables of concrete param values, fully covering the common cases the description promises.

5 / 5

Workflow Clarity

Workflows are clearly sequenced with validation checkpoints and feedback loops: check `state()` readiness before calling, interpret the resolved window line, verify "views grew and seen is near 0 s" when narrowing, and error recovery paths such as halving labels on torch.OutOfMemoryError and reporting (not restarting) a down stack.

5 / 5

Progressive Disclosure

The body is well-sectioned with clear headers and one clearly-signaled one-level-deep pointer ("Tuning params offline on a recording... is in `README.md`"), and no bundle files exist so self-containment is appropriate. It stops short of a 5 because reference material like the LocalizeParams field table and the rig failure guide are inlined and could be split into a references/ file to slim the overview.

4 / 5

Total

18

/

20

Passed

Description

92%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong description: it states what the skill does through six concrete actions, names the tools involved, and gives an explicit 'Use whenever' trigger anchored to specific blueprint names. The only weakness is the absence of a few natural synonyms for 'locate' that users might phrase differently.

DimensionReasoningScore

Specificity

The description lists six concrete actions ("probe the machine, build the time window, watch labels on a schedule, narrow an object's position, change params per call, and tell where a failure comes from") and names the exact tools (`state`, `localize`), matching the comprehensive-coverage anchor. It does not fit the 4 anchor because action coverage has no notable gaps.

5 / 5

Completeness

Both parts are explicit: the 'what' ("Locate objects with a running PerceptionLocalize stack over MCP") and the 'when' with concrete trigger phrases ("Use whenever a blueprint with PerceptionLocalize and McpServer runs, such as go2-localize or xarm-localize-live"), matching the top anchor exactly.

5 / 5

Trigger Term Quality

Good keyword coverage within the niche: "Locate objects", "PerceptionLocalize", "MCP", `state`, `localize`, and the exact blueprint names "go2-localize" and "xarm-localize-live" that a user of this stack would naturally say. It falls short of the 5 anchor because common synonyms such as "find", "detect", or "where is" are absent, but it is well above the 3 anchor's missing-variations level.

4 / 5

Distinctiveness Conflict Risk

The description names a specific stack (PerceptionLocalize over MCP), its tools, and the concrete blueprints that run it, giving it a clear niche with minimal overlap risk against generic localization or object-detection skills.

5 / 5

Total

19

/

20

Passed

Validation

100%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository

dimensionalOS/dimos

Path

dimos/perception/localize/SKILL.md

Reviewed

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.