CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

codex-issue-waves

Run a batch of GitHub issues through codex exec in isolated git worktrees as parallel autonomous PRs, then manage the review and correction waves until merge. Use when the user gives a list of issue numbers (≥ 2) and asks to "spawn codex" / "dispatch codex" / "have codex work on" / "manage the PRs" / "process feedback" / "get them merged" for those issues, or when the user asks for multi-issue parallel delegation to codex. Not for single-issue wave-driven delegation (use codex-task-waves), single-issue one-shot dispatch (use invoking-codex-exec), or implementation without delegation (use /pr or direct implementation).

Quality

97%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Low

Low-risk findings worth noting

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.

A strong, highly actionable orchestration skill: exact commands and scripts, a clearly sequenced multi-phase workflow with validation, feedback loops, and escalation limits, and clean progressive disclosure to verified bundle files. The only flaw is deliberate repetition of the PO/worker boundary and status-vs-content rules across three sections, which costs it the top conciseness score.

DimensionReasoningScore

Conciseness

Efficient and workflow-specific — no space is spent explaining git/CI/codex concepts Claude already knows, and templates, shell recipes, and pitfalls are offloaded to bundle files. However, the orchestrator-boundary rule ("Claude never edits source files, never reads diffs for judgment") and the status-vs-content distinction are each repeated about three times (orchestration model, review wave, and Key pitfalls), which is trimmable repetition. This fits anchor 4 ('minor instances of over-explanation that could be trimmed') better than 3, since the rest is dense with non-obvious operational knowledge, and better than 5, since 'every token earns its place' is violated by the deliberate repetition.

4 / 5

Actionability

Fully executable guidance throughout: exact commands (`gh pr view <N> --json statusCheckRollup,mergeable,mergeStateStatus,headRefOid`, `git diff origin/main...origin/<branch> --stat`), concrete scripts with argument signatures (`scripts/merge_and_cleanup.sh <PR> <worktree-path> <branch>`, `scripts/wait_for_ci.sh <PR>`), exact worktree/branch naming conventions, a precise log-line format with field semantics, and a copy-paste archival shell snippet. Prompt composition steps delegate to a real, verified template file. Matches the copy-paste-ready top anchor.

5 / 5

Workflow Clarity

Clear four-phase sequence (dispatch → monitor → review wave → correction wave → merge + cleanup) with explicit validation checkpoints at every phase: PR existence/CI checks with exact commands, malformed-JSON handling, read-only enforcement verification, verification of log-vs-committed conflict markers via `gh pr diff`, and a defined feedback loop (corrector → re-dispatch reviewer) with an escalation cap after 3 cycles. The conflict-triage checklist and Success criteria audit section complete the pattern. As a batch operation it fully satisfies the validation requirement, so no cap applies.

5 / 5

Progressive Disclosure

The body is a well-organized orchestration overview that offloads detail to real, verified, one-level-deep bundle files with clear signals: "See `references/prompt-template.md` for the shape", "See `references/common-commands.md` for the exact shell recipes", and "See `references/pitfalls.md` for the full list. The top items:". All five referenced files (three references, two executable scripts) exist with substantive matching content, and no reference chain nests deeper than one level. The inlined reviewer checklist is appropriately inlined since it must be pasted verbatim into reviewer prompts.

5 / 5

Total

19

/

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.

The description is exemplary: concrete multi-action capability statement, explicit 'Use when' trigger guidance with natural quoted phrases, a quantified boundary (≥ 2 issues), and explicit disambiguation against named sibling skills. It fully satisfies the what/when/distinctness requirements with no fluff despite its length.

DimensionReasoningScore

Specificity

Lists multiple concrete actions covering the full lifecycle — "Run a batch of GitHub issues through codex exec in isolated git worktrees as parallel autonomous PRs", "manage the review and correction waves until merge" — naming the exact mechanism (codex exec, worktrees, PRs) rather than generic language. Coverage is comprehensive across dispatch, review, correction, and merge, matching the top anchor; score 4 would require minor gaps in coverage, which are absent.

5 / 5

Completeness

Explicitly answers both questions: what ("Run a batch of GitHub issues through codex exec in isolated git worktrees as parallel autonomous PRs, then manage the review and correction waves until merge") and when ("Use when the user gives a list of issue numbers (≥ 2) and asks to...") with concrete trigger phrases. Matches the anchor 5 example structure exactly; anchor 4 would require a less explicit or less specific 'when' clause.

5 / 5

Trigger Term Quality

Includes extensive natural quoted trigger phrases users would actually say — "spawn codex" / "dispatch codex" / "have codex work on" / "manage the PRs" / "process feedback" / "get them merged" — plus the descriptive variant "multi-issue parallel delegation to codex". This is comprehensive coverage with synonyms and variations; there is no jargon-only or generic phrasing that would drop it to 4.

5 / 5

Distinctiveness Conflict Risk

Has a clear niche (multi-issue, ≥2, codex delegation) with an explicit negative scope clause naming sibling skills: "Not for single-issue wave-driven delegation (use codex-task-waves), single-issue one-shot dispatch (use invoking-codex-exec), or implementation without delegation (use /pr or direct implementation)". This active disambiguation minimizes conflict risk beyond the 'minor overlap' of anchor 4.

5 / 5

Total

20

/

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

ddnetters/homelab-agent-skills

Path

codex-issue-waves/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.