CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

hivemind

Orchestrate free opencode workers from Claude Code to cut token costs. Use when delegating grunt work to a single worker or a parallel swarm (scout/coder/tester) with worktree isolation, benchmarking against opencode, or when the user says "spawn a worker", "swarm", "delegate to opencode", or "/oc".

75

Quality

94%

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

92%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.

Excellent operational content: exact executable commands, sequenced flows with explicit validation and error-recovery loops, and a well-structured bundle whose referenced files all exist. The single deduction is duplicated retry/fallback guidance repeated across three sections, which inflates token cost without adding information.

Suggestions

State the retry-once policy in one authoritative place (e.g., the fallback ladder) and have the 'Free models are not always up' section and the anti-patterns table reference it instead of restating it.

Merge the overlapping escalation rules — the swarm flow's 'escalate to your own Sonnet … fails twice' and the fallback ladder's step 2 'orchestrator performs that subtask inline' — into a single, consistent escalation path.

Consider moving the Windows-specific resolver notes ('Windows notes (hard-won)') into a short reference file, keeping SKILL.md to the one-line OPENCODE_GIT_BASH_PATH prerequisite.

DimensionReasoningScore

Conciseness

The body is dense and table-driven with no explanations of concepts Claude already knows, but retry-once guidance is stated redundantly in three places ("Free models are not always up", the "Fallback ladder", and the "Retrying a failing worker indefinitely" anti-pattern row), which is trimmable duplication. This fits anchor 4 ('minor instances that could be trimmed') rather than anchor 5, where every token would earn its place.

4 / 5

Actionability

Invocations are copy-paste ready with exact flags, return shapes, and env vars: "node \"<skill-dir>\\scripts\\oc-worker.mjs\" [--agent scout|coder|tester] [--dir <path>] …", "git worktree add ../<repo>-wt-N -b swarm/N", and "node scripts/oc-verify.mjs --dir <same dir the worker saw> --claims findings.json --run <id>". Specific commands cover the common cases, matching the anchor-5 'fully executable' example.

5 / 5

Workflow Clarity

The swarm flow is a numbered 1-6 sequence with explicit validation checkpoints ("Review every diff yourself", "YOU are the only merger", "run tests"), and the fallback ladder plus the oc-verify stage provide explicit error-recovery feedback loops for batch operations. This matches anchor 5's 'clear sequence with explicit validation steps; feedback loops for error recovery' — no checkpoint is missing or implicit.

5 / 5

Progressive Disclosure

The body is an overview with a components table routing to every bundle file (all verified to exist on disk), with command definitions split into assets/commands/ and worker personas into assets/agents/ — one level deep and clearly signaled. Nothing that belongs in a separate file is inlined, matching anchor 5 rather than anchor 4's 'minor organization gaps'.

5 / 5

Total

19

/

20

Passed

Description

95%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: concrete actions, explicit 'Use when' triggers with verbatim user phrases, and a distinct niche. The only weakness is that the capability list stops short of the verification and fleet-review flows the skill also provides, leaving minor coverage gaps.

DimensionReasoningScore

Specificity

"Orchestrate free opencode workers from Claude Code to cut token costs" and "delegating grunt work to a single worker or a parallel swarm (scout/coder/tester) with worktree isolation, benchmarking against opencode" list several concrete actions. It is not 5 because coverage has minor gaps (verification, review panels, and the other fleet commands are not mentioned), and well above the 1-2 actions of anchor 3.

4 / 5

Completeness

It explicitly answers both what ("Orchestrate free opencode workers from Claude Code to cut token costs") and when ("Use when delegating grunt work to a single worker or a parallel swarm … or when the user says …") with concrete trigger phrases, exactly matching the anchor-5 example. A weaker 'when' would cap it at 4, but this one is fully explicit.

5 / 5

Trigger Term Quality

The description quotes natural user phrases verbatim — "spawn a worker", "swarm", "delegate to opencode", or "/oc" — alongside "grunt work", "token costs", and "benchmarking". Coverage of natural terms including slash-command synonyms is comprehensive, matching the anchor-5 example.

5 / 5

Distinctiveness Conflict Risk

It occupies a clear niche (opencode worker orchestration) with distinct, tool-specific triggers ("delegate to opencode", "/oc", "spawn a worker") that are unlikely to fire for other skills, matching anchor 5. It is more distinct than anchor 4's 'minor overlap risk with closely related skills'.

5 / 5

Total

19

/

20

Passed

Validation

93%

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

Validation — 15 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

referenced_paths_exist

Referenced path issues: 2 deeper-than-1-level

Warning

Total

15

/

16

Passed

Repository

alirezarezvani/claude-skills

Path

engineering/hivemind/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.