tdg-personal/dmux-workflows
Multi-agent orchestration using dmux (tmux pane manager for AI agents). Patterns for parallel agent workflows across Claude Code, Codex, OpenCode, and other harnesses. Use when running multiple agent sessions in parallel or coordinating multi-agent development workflows.
75
75%
Does it follow best practices?
Impact
Pending
No eval scenarios have been run
Passed
No known issues
Quality
Discovery
89%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
This is a solid description with clear 'when' triggers and a distinctive niche. The main weakness is that the specific capabilities could be more concrete — listing actual actions like spawning panes, broadcasting commands, or monitoring agent output would strengthen it. The trigger term coverage is strong with multiple natural keywords users would employ.
Suggestions
Add more concrete actions beyond 'patterns for parallel agent workflows' — e.g., 'spawn tmux panes, broadcast commands to agents, monitor parallel sessions, coordinate task splitting across agents'
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Names the domain (multi-agent orchestration) and mentions some actions (parallel agent workflows, coordinating workflows), but doesn't list multiple specific concrete actions like 'spawn panes', 'broadcast commands', 'monitor agent output', etc. | 2 / 3 |
Completeness | Clearly answers both 'what' (multi-agent orchestration using dmux for parallel agent workflows) and 'when' (explicit 'Use when running multiple agent sessions in parallel or coordinating multi-agent development workflows'). | 3 / 3 |
Trigger Term Quality | Good coverage of natural terms users would say: 'dmux', 'tmux', 'multi-agent', 'parallel', 'Claude Code', 'Codex', 'OpenCode', 'agent sessions', 'orchestration'. These are terms users working with multi-agent setups would naturally use. | 3 / 3 |
Distinctiveness Conflict Risk | Very distinct niche — dmux/tmux pane management for AI agents is highly specific and unlikely to conflict with other skills. The mention of specific tools (dmux, tmux, Claude Code, Codex, OpenCode) creates clear differentiation. | 3 / 3 |
Total | 11 / 12 Passed |
Implementation
50%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This skill provides a solid overview of dmux-based multi-agent orchestration with useful workflow patterns and practical tips. However, it's longer than necessary — the ECC Helper section alone adds ~40 lines of JSON examples that could be in a separate file. The core weakness is that workflow patterns lack validation checkpoints (especially around merging) and the guidance is more illustrative than directly executable.
Suggestions
Add explicit validation/verification steps to workflow patterns, especially after merge operations (e.g., 'After merging, run tests to verify no conflicts: `npm test`' or 'Review merged output before committing').
Move the ECC Helper section and its JSON examples into a separate reference file (e.g., ECC_ORCHESTRATION.md) and link to it from the main skill.
Make the Quick Start more actionable by showing the actual dmux CLI commands rather than describing UI keypresses in comments (e.g., `dmux new 'Implement auth middleware'` if such a CLI exists, or clarify the exact interaction model).
Trim the 'What is dmux' section and complementary tools table — Claude doesn't need the tool explained, just the usage patterns.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is mostly efficient but includes some unnecessary content. The 'What is dmux' section explains things Claude could infer, the complementary tools table adds marginal value, and the ECC Helper section with its lengthy JSON examples adds significant bulk. The pattern descriptions could be tighter. | 2 / 3 |
Actionability | The workflow patterns provide concrete prompt examples but are illustrative rather than truly executable — they show what to type into dmux panes as comments/strings rather than providing copy-paste-ready commands. The git worktree section and ECC helper section have executable commands, but the core dmux interaction patterns rely on UI actions (pressing 'n', 'm') that can't be scripted directly. | 2 / 3 |
Workflow Clarity | Patterns show clear sequences (research then implement, parallelize then merge), but lack explicit validation checkpoints. There's no guidance on verifying merge results, checking for conflicts after merging panes, or confirming agent completion before proceeding. The merge step — a potentially destructive operation — has no validation loop. | 2 / 3 |
Progressive Disclosure | The content is reasonably structured with clear sections, but it's monolithic — the ECC Helper section with its JSON examples and the detailed worktree integration could be split into separate reference files. The complementary tools table and troubleshooting could also be externalized. No references to separate files for advanced content. | 2 / 3 |
Total | 8 / 12 Passed |
Validation
90%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 10 / 11 Passed
Validation for skill structure
| Criteria | Description | Result |
|---|---|---|
frontmatter_unknown_keys | Unknown frontmatter key(s) found; consider removing or moving to metadata | Warning |
Total | 10 / 11 Passed | |
Reviewed
Table of Contents