CtrlK
BlogDocsLog inGet started
Tessl Logo

doca-common

Use this skill whenever the user is doing hands-on DOCA programming on a BlueField DPU or ConnectX NIC and needs the foundation primitives every per-library context rests on — walking the doca_ctx lifecycle, discovering doca_dev / doca_devinfo and gating on doca_*_cap_* before trusting a feature, wiring doca_mmap / doca_buf_inventory / doca_buf for zero-copy I/O across libraries, driving doca_pe for completions, or DOCA Log's two-tier (--sdk-log-level vs app-side) model. Trigger even when the user does not say "DOCA Common" — typical implicit phrasings include "my tasks submit but nothing completes", "DOCA_ERROR_BAD_STATE from doca_ctx_start", "--sdk-log-level does nothing for my DOCA_LOG_DBG lines", "share a buf between doca_dma and doca_rdma", or "crashes far from the offending line". Refuse and route elsewhere for per-library questions in isolation (load doca-flow / doca-rdma / doca-eth alongside), installing DOCA (doca-setup), or doc lookup (doca-public-knowledge-map).

68

Quality

83%

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

67%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.

A well-architected thin-loader: precise routing to companion files and sibling skills with concrete commands and paths. The main weakness is token-efficiency — the primitive set is enumerated many times — and the referenced companion files are not present in the bundle.

Suggestions

Consolidate the primitive enumeration (log/buf/ctx/dev/progress engine) into one canonical list; it currently repeats across 'Where to start', 'Example questions', 'When to load', 'What this skill provides', and 'Loading order'.

Replace the six prose 'Example questions' blocks with a compact question-class → file/section table to halve the tokens while preserving the routing signal.

Ship CAPABILITIES.md and TASKS.md in the skill bundle (or confirm their resolved path); the body routes to them but neither is present alongside SKILL.md or under ./references/.

DimensionReasoningScore

Conciseness

No basic-concept padding (it assumes Claude knows what a DPU/library is), but the same primitive enumeration (log/buf/ctx/dev/progress engine) is repeated 5–6 times across 'Where to start', 'Example questions', 'When to load', 'What this skill provides', and 'Loading order', which could be tightened to one canonical list. It is not a 2 because the verbosity is structural repetition rather than explaining things Claude already knows.

3 / 5

Actionability

Concrete, specific routing — exact file+section anchors (CAPABILITIES.md ## ctx, TASKS.md ## configure), a real command ('pkg-config --modversion doca-common'), and sample paths (/opt/mellanox/doca/samples/<library>/<sample>/*_main.c). It is a deliberate no-code loader, so the absence of code is by design; the minor gap is that most guidance is routing rather than directly executable steps.

4 / 5

Workflow Clarity

A clear 3-step Loading Order (read SKILL.md → CAPABILITIES.md → TASKS.md) and per-question-class routing to specific anchors. No validation checkpoints are needed because this is a pure loader (no destructive/batch operations to guard), so the simple-skill rationale applies; not a 5 only because the 'Example questions' mappings are prose-heavy rather than a crisp sequenced checklist.

4 / 5

Progressive Disclosure

Well-signaled, one-level-deep references with markdown links and anchors to two companion files plus cross-links to sibling skills — structurally a 5. It drops to 4 because the referenced companion files (CAPABILITIES.md, TASKS.md) are not present in the skill bundle (no references/scripts/assets dirs), so the disclosure points to detail that is not actually shipped here.

4 / 5

Total

15

/

20

Passed

Description

100%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.

An exemplar description: concrete capabilities, natural symptom-based triggers, explicit what-and-when, and crisp routing boundaries against sibling DOCA skills. Third-person voice is maintained throughout.

DimensionReasoningScore

Specificity

Lists multiple concrete actions — 'walking the doca_ctx lifecycle', 'discovering doca_dev / doca_devinfo and gating on doca_*_cap_*', 'wiring doca_mmap / doca_buf_inventory / doca_buf for zero-copy I/O', 'driving doca_pe for completions', and the two-tier log model — giving comprehensive, non-abstract coverage.

5 / 5

Completeness

Explicitly answers both what (the foundation primitives enumerated) and when ('Use this skill whenever...', plus 'Trigger even when the user does not say DOCA Common — typical implicit phrasings include...'), and adds explicit refuse/route guidance.

5 / 5

Trigger Term Quality

Gives natural symptom phrasings a user would actually say ('my tasks submit but nothing completes', 'DOCA_ERROR_BAD_STATE from doca_ctx_start', 'share a buf between doca_dma and doca_rdma', 'crashes far from the offending line'), which is exactly the 'would a user naturally say this' test; not just technical jargon.

5 / 5

Distinctiveness Conflict Risk

Clear niche (doca-common foundation primitives under every per-library context) with explicit routing away from per-library questions in isolation, install (doca-setup), and doc lookup (doca-public-knowledge-map), minimizing conflict with sibling skills.

5 / 5

Total

20

/

20

Passed

Validation

87%

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

Validation14 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

metadata_version

'metadata.version' is missing

Warning

relative_links

Relative link issues: 20 missing, 17 suspicious

Warning

Total

14

/

16

Passed

Repository
NVIDIA/skills
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.