דילוג לתוכן
New

DeepSeek V4.1 Flash free for 7 days: 10M tokens a day after a top-up of $10 or more. Until Nov 30.

How it works
trailofbits/code-improverWeb ו-bug bounty

code-improver

⁨Runs an autonomous review-and-fix improvement loop over any code target — a skill, plugin, module, or directory — using a reviewer the user names: any installed skill or agent. Keeps a cross-round findings ledger, escalates when fixes stop converging, and guards scope mechanically. Use when asked to⁩

v2026.11.12-149d7b23 התקנות⁦+3⁩ ב-7 ימים8.2K‏ כוכביםCC-BY-SA-4.0סונכרן לפני 22 שעות

התקנה

עדיין אין לכם את ה-CLI של NotHoly? התקינו אותו קודם.

$ curl -fsSL https://notholy.com/install.sh | sh
$ notholy skills install trailofbits/code-improver

נכתב אל ⁦~/.notholy/skills⁩ אחרי בדיקת החתימה וה-SHA-256.

חתימהחתום, ed25519
QArSK/fCzb…Gf401cAg==ה-CLI בודק את ה-SHA-256 ואת החתימה לפני שהוא כותב קובץ כלשהו.
חבילהc0e36186…30a2fd47
c0e361864e72…5a8930a2fd47קובץ אחד · ⁦3.0 KB⁩
סריקהנמוךאין ממצאים
הסריקה לא מצאה בגרסה הזו שום דבר שצריך לסמן.לסריקה המלאה
מקורgithub.com/trailofbits/skills/tree/main/plugins/code-improver/skills/code-improver@149d7b2

הפריט הזה עדיין לא נכתב בעברית, ולכן הוא מוצג בשפה אחרת: English.

Improve any code target by running /code-improver:improve — a dynamic workflow that loops the named reviewer and a fixer subagent until a review reports zero critical/major findings, then strips its own residue. The loop, its ledger, and its guards live in the workflow; this skill collects the three inputs the generic entry requires and relays the outcome.

Starting the loop

The user provided: $ARGUMENTS (if empty, take the details from the conversation).

1. Collect the three required inputs — no guessing

  1. Target: the absolute path to the directory under improvement. Resolve relative paths against the working directory; verify the directory exists.
  2. Reviewer: the installed skill or agent that performs every review. The user must name it — there is no default and no bundled reviewer. Determine the kind:
    • a namespaced agent (e.g. plugin-dev:skill-reviewer) → "kind": "agent"
    • an installed skill (e.g. pr-review-toolkit:review-pr) → "kind": "skill" If the name could be either, check the session's skill listing; if still ambiguous, ask the user. If no reviewer was named, ask — do not pick one.
  3. Scope: repo-relative globs the loop may touch. The generic entry requires it explicitly; if the user did not give one, propose the target directory (<repo-relative-target>/**) and confirm before launching.

2. Resolve the loop script

Resolve improve.js from this SKILL.md's installed location to an absolute path and confirm the file exists. Its plugin directory is pluginRoot. Use this exact plugin copy; if the file is missing, stop.

3. Invoke the workflow

Run it with the Workflow tool, {scriptPath: "<the path from step 2>", args: {...}}:

json
{
  "target": "<absolute target path>",
  "reviewer": { "kind": "agent|skill", "name": "<namespaced-name>", "notes": "<what the reviewer should know about the target>" },
  "scope": ["<repo-relative-glob>/**"],
  "pluginRoot": "<the plugin directory from step 2>",
  "maxRounds": 5
}
  • maxRounds only if the user asked for a different cap.
  • pluginRoot is the resolved plugin directory from step 2, for the metrics collector.
  • finalize ({"version_bump": bool, "narration_strip": bool, "docs_pass": bool}) only to override the defaults: version bump when the target sits inside a plugin, narration strip and docs pass always.
  • decision only on continuation (below).

The workflow runs in the background and needs no babysitting: it reviews, fixes, re-reviews, checks scope after every fix round, and can only complete on a clean review. It never commits; all changes stay in the working tree.

If the Workflow tool is unavailable or denied, stop and say so. Do not improvise the loop inline with direct edits — the ledger, scope guard, and escalation guarantees live in the workflow, and an inline imitation has none of them.

If the result is halted: "reviewer-unavailable", relay it and stop. The named reviewer is not installed in this session; tell the user which plugin provides it and re-run after installing. Do not review the target yourself.

Do not end your turn while the loop is running. The Workflow tool returns a task id immediately; the result comes later. In an interactive session the completion notification re-invokes you — wait for it. In a non-interactive run (scripted, CI, eval) there is no later turn: stopping abandons the loop mid-round, so after launching, poll the task (TaskOutput with the returned task id, or sleep-and-recheck) until it completes, then relay the result. A session that answers "the loop is running, I'll report later" has lost the run.

Relaying the result

The workflow returns a structured result. Report it honestly — the distinctions matter:

  • converged: true — the last action was a review with zero critical/major findings. Report rounds used, remaining minor findings (open_minor_count), and the artifact paths (ledger_path, metrics).
  • capped: true — the fix budget ran out and the FINAL review still found blocking issues. Say plainly: capped, NOT converged, and list open_blocking. Do not present this as success.
  • escalation — the loop detected it was not converging (recurring findings, non-decreasing counts, or a fix relocating a problem). Relay the escalation message and finding ids to the user: this needs a design decision, not more rounds.
  • halted — a guard fired (scope violation, unregistered new files, a dead or unavailable reviewer, or a finalize pass whose own edits failed the check that follows it). Relay the paths in violations/new_untracked_files, the sites in finalize_regressions, and the notes.
  • notes always travel with the result — surface them; they include loud warnings such as "a git repository was initialized".

Continuing after an escalation

The loop stops on escalation by design. When the user decides, start a fresh run with the same target and reviewer plus:

json
{ "decision": "<the user's ruling, verbatim>" }

The new run reloads the on-disk ledger, so every finding, rejection, and verdict carries over — rounds restart, re-derivation does not.

To stop a running loop, stop the workflow task (TaskStop); the ledger on disk is current to the last round and a re-run resumes from it.

When NOT to use

  • A Claude Code skill: use the skill-improver entry — it wires the right reviewer
  • A branch / pull request: use the pr-improver entry — it derives scope from the diff
  • One-time review: dispatch the reviewer directly; the loop's value is iteration
  • Quick single fixes: edit the file directly

הקבצים בגרסה הזו

הרשאות

הרשאההוצהרנמצא בסריקה
קורא קבציםכןכן
כותב קבציםלאלא
מריץ פקודות shellכןכן
משתמש ברשתלאלא
משתמש במחבריםלאלא
שולט בדפדפןלאלא

ממצאי הסריקה

נבדק ב-13 בנוב׳ 2026

הסריקה לא מצאה בגרסה הזו שום דבר שצריך לסמן.

סריקה נקייה מורידה את הסיכון, אבל לא מוכיחה שהחבילה בטוחה. כדאי לקרוא מה מתקינים.

גרסהפורסמהCommitSHA-256חתימהשינוי בהרשאות
v2026.11.12-149d7b213 בנוב׳ 2026149d7b2c0e36186חתומה–

שאלו כל מה שהסקופ מאפשר.

יוצרים חשבון, מוסיפים קרדיטים ושואלים את השאלה הראשונה בצ׳אט או בטרמינל.

  1. 1/3

    יצירת חשבון

    פרטי או חברה. אפשר לעבור לחשבון חברה אחר כך בלי לאבד יתרה או מפתחות.

    פרטיחברה
  2. 2/3

    הוספת קרדיטים

    בכרטיס אשראי. תשלום לפי טוקן, בלי מנוי.

    $5.00טעינה מינימלית
  3. 3/3

    שאלה ראשונה

    בצ׳אט, או להתקין את ה-CLI ולהריץ notholy login בטרמינל.

    $ notholy login