CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

add-opencode

Use OpenCode as an agent provider. OpenRouter, OpenAI, Google, DeepSeek, etc. via OpenCode config — not the Anthropic Agent SDK. Per group via `ncl groups config update --provider opencode`; host passes OPENCODE_* and XDG mount when spawning containers.

Quality

75%

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

Fix and improve this skill with Tessl

tessl review fix ./.claude/skills/add-opencode/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

81%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 highly actionable, well-validated operational manual: every step is executable, gated by explicit checks, and paired with error-recovery guidance. The main weaknesses are conciseness (repeated refresh/pin semantics and run-on edge-case prose) and progressive disclosure (dangling reference targets in the provided bundle and heavy inlining that a split into reference files would relieve).

Suggestions

Consolidate the repeated '1.18.25' pins and refresh semantics into one stated place (e.g. a short 'Pins and refresh' note) instead of restating them in Install, Authenticate, and the native-test section.

Break the opening edge-case paragraphs into short bulleted notes so each install/auth caveat is scannable rather than buried in run-on prose.

Move the file copy manifest and env-variable reference into a linked reference file (and ship ARCHITECTURE.md/REMOVE.md alongside the skill) so SKILL.md reads as an overview with verifiable one-level-deep references.

DimensionReasoningScore

Conciseness

Almost every line is system-specific operational fact with no padding or explanations of known concepts, but the opening paragraphs are dense run-on edge-case dumps, refresh/pin semantics are repeated across the Install and Authenticate sections, and the exact version '1.18.25' appears inline six times instead of being consolidated. Not 4 because tightening requires more than minor trims; not 2 because nothing is concept-explaining filler.

3 / 5

Actionability

Fully executable throughout: copy-paste commands ('pnpm exec tsx setup/index.ts --step provider-auth opencode', 'ncl groups config update --id <group-id> --provider opencode'), a complete runnable compatibility-check one-liner, an exact file copy manifest, pinned dependency declarations, and a runnable native-test recipe with 'OPENCODE_TEST_BINARY'. Covers the common install/auth/model cases completely.

5 / 5

Workflow Clarity

Clear Install → Authenticate → Recover login → Change model → Backend defaults sequence with explicit validation and feedback loops: the compatibility predicate 'guards every subsequent step', tests 'must pass before rebuilding the agent image', 'Send a message and verify a reply, then send a second message', and error recovery for vault/duplicate-name failures and expired logins ('--reauth', 'Retry the failed request').

5 / 5

Progressive Disclosure

Sections are well organized and the outbound links ([ARCHITECTURE.md], [REMOVE.md]) are one level deep and clearly signaled, but the provided bundle contains no such files (no references/, payload/, or docs directories), so the reference targets cannot be verified, and ~345 lines of operational detail (copy manifest, env-var reference, auth recovery) are fully inlined. Not 5 because content is not split into an overview plus detailed reference files; not 3 because structure and signaling are good.

4 / 5

Total

17

/

20

Passed

Description

61%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.

The description is specific and distinctive, with concrete commands and backend names, but it lacks any explicit 'when to use' trigger guidance, which is its most heavily weighted gap. It also undersells the skill's actual scope (install, authentication, model management, login recovery). Adding an explicit trigger clause and naming the core workflows would lift it substantially.

Suggestions

Add an explicit trigger clause, e.g. 'Use when setting up, authenticating, or changing the model for an OpenCode-backed agent group in NanoClaw.'

Name the concrete workflows the skill covers (install payload, authenticate via ChatGPT/API key, rotate/recover logins, change the default model) so the 'what' is comprehensive rather than a single action.

Include one or two natural user phrasings such as 'sign in to OpenCode' or 'switch a group to OpenCode' to broaden trigger-term coverage.

DimensionReasoningScore

Specificity

Names the domain ('Use OpenCode as an agent provider') with one concrete command ('ncl groups config update --provider opencode') and a backend list, but never mentions the skill's main workflows (install, authenticate, model selection, login recovery). It is not 4 because the actions described are minimal configuration mechanics rather than several specific capabilities.

3 / 5

Completeness

The 'what' is clear ('Use OpenCode as an agent provider') but there is no 'Use when...' clause or equivalent trigger guidance, which caps completeness at 3. Not 4 because the 'when' is entirely absent rather than merely implicit.

3 / 5

Trigger Term Quality

Includes natural product terms users would say — 'OpenCode', 'OpenRouter', 'OpenAI', 'Google', 'DeepSeek' — but omits natural phrasings for the covered workflows such as signing in, authenticating, choosing a model, or provider setup. Not 5 because coverage beyond product names is thin.

4 / 5

Distinctiveness Conflict Risk

The 'not the Anthropic Agent SDK' contrast plus the specific backend names and per-group config command carve a distinct niche with minimal conflict risk. Not 5 because 'agent provider' phrasing could overlap with sibling provider-setup skills in the same ecosystem.

4 / 5

Total

14

/

20

Passed

Validation

81%

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

Validation — 13 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

metadata_version

'metadata.version' is missing

Warning

relative_links

Relative link issues: 2 missing

Warning

referenced_paths_exist

Referenced path issues: 24 missing

Warning

Total

13

/

16

Passed

Repository

nanocoai/nanoclaw

Path

.claude/skills/add-opencode/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.