| name | ts-review-ultra |
| description | Use when you want a full review pass of TypeScript code. Triggers on "ultra-review src/modules/foo", "run all the ts reviews on this module", "give me everything you've got on this file". Fans out ts-review-style, ts-review-nit, and ts-review-skeptic in parallel via the Agent tool, deduplicates findings, and returns one combined report. Does not write code. |
Fan out the three ts-review skills in parallel against a target directory, then merge.
Input
The user gives you a target directory or file (src/modules/foo, a path, or a project reference). If the user has a claim to scrutinize (the ts-review-nit phase 1 input), capture it too. If the directory is missing, ask. If the claim is missing, proceed without it and tell ts-review-nit to skip phase 1.
How to fan out
Spawn three Agent tool calls in a single message (parallel), using the dedicated agent types:
- ts-review-style (
subagent_type: ts-review-style): apply TypeScript style and judgment conventions.
- ts-review-nit (
subagent_type: ts-review-nit): phase 1 (verify claim) + phase 2 (nit exported API). If no claim, tell it to skip phase 1.
- ts-review-skeptic (
subagent_type: ts-review-skeptic): phase 1 (bug hunt) + phase 2 (deletion sweep). No praise.
Prompt each agent with the target directory and any claim. They already have their instructions built in; just tell them what to review.
ts-review-legibility: only run if the user supplied a list of inspected false-positive findings. This one stays as a skill (invoked via the Skill tool) because it needs user-provided input. If skipped, note it in the merged report.
Model selection
The Agent tool accepts an optional model parameter (sonnet / opus / haiku) that overrides the subagent's default per call. Support both a default and a per-invocation override:
- Default: no
model override. Subagents inherit the parent session's model. Keeps cost predictable.
- Per-invocation override. Parse a model hint from the user's prompt, e.g. "ts-review-ultra src/modules/foo with opus", "...on sonnet", "...using haiku". If a model is named, pass
model: <name> on every fanned-out Agent call. If the user gives a per-lens hint ("opus for skeptic, sonnet for the rest"), honor it.
- Recognized phrasing: literal model name (
opus / sonnet / haiku), optionally with on / with / using / via. Reject ambiguous hints by asking which model the user means.
Merging the findings
Each subagent returns a list. Merge into one report:
- Group by skill. Keep the three sections separate. The lenses are different and collapsing them loses signal.
- Dedupe by
file:line within a section. If style and nit both flag the same line for related reasons, keep both (they're different lenses). Only dedupe exact restatements.
- Per finding, include a "why" line. Don't just say what's wrong. Explain what goes wrong in practice if the pattern isn't fixed. The reader should learn the principle, not just the fix. Example: "Missing
AbortSignal on a network call. If a Temporal activity is cancelled, this HTTP request keeps running and consuming resources until it times out on its own."
- Combined severity table at the end, drawn from all three sections. Promote any finding the subagent marked high to the top.
- Top 1-3 to action first across all skills, opinionated. Don't list ties.
If ts-review-nit's phase 1 pushed back on the claim, surface that before the merged findings. The user needs to redirect before the rest of the report matters.
Domain-specific false positives
When the target contains Temporal workflows or activities, include this context in each subagent prompt:
Temporal workflow files have constraints that override normal TypeScript conventions:
- Exported Input/Output interfaces on workflows are the public contract for
client.workflow.start() callers. Do not flag these as "only used by specs" or recommend dropping export.
- Workflow functions must be standalone exported functions, not class methods. Do not flag mixed function/method styles between workflows and activities.
- Duplicate
defineSearchAttributeKey calls across workflow files may be required because each workflow is compiled into an isolated bundle. Verify bundle boundaries before recommending extraction.
String(err) in workflow log calls is sometimes necessary because Temporal's workflow sandbox restricts what can be serialized. Flag it but mark as needs-verification.- Sequential
startChild in a loop is intentional (avoids thundering-herd). Do not flag as "should use Promise.all".
Gotchas
- Agents don't inherit conversation context. Pass the target dir, the claim (if any), and any constraints in the spawn prompt.
- 3x the token cost of a single review. The win is parallelism and keeping intermediate searches out of the main context.
What this skill does not do
- Write code or apply fixes. Output is a merged findings report; the user picks what to fix.
- Run
ts-review-optimize. That skill is invoke-explicitly-only and is not part of the standard review pass.
