[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7CH72V3GPK94HH8FFMKGFYN":3,"skill-same:01M7CH72V3GPK94HH8FFMKGFYN":46},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":17,"publishedAt":18,"linkOnly":19,"markdownContent":20,"markdownBody":21,"frontMatter":22,"refName":25,"commitSha":26,"skillDirectory":27,"forksCount":28,"repositoryHtmlUrl":29,"repositoryHomepageUrl":6,"ownerHtmlUrl":30,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":31,"install":32,"resources":45},"01M7CH72V3GPK94HH8FFMKGFYN","code-improver",null,"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 'improve this code until review passes', 'run an improvement loop with \u003Creviewer>', or to iterate review-and-fix with a specific reviewer. For skills prefer the skill-improver entry; for a branch prefer pr-improver.","bongkage\u002Ftrailofbits-skills","bongkage","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F101378796?v=4","01M7CH419QTJ16DNJ33P0WCWBE","01M7CH41A9NWK8D2JBRK9F0M2C",1,"CC-BY-SA-4.0","https:\u002F\u002Fgithub.com\u002Fbongkage\u002Ftrailofbits-skills\u002Fblob\u002Fmain\u002Fplugins\u002Fcode-improver\u002Fskills\u002Fcode-improver\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: code-improver\ndescription: \"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 'improve this code until review passes', 'run an improvement loop with \u003Creviewer>', or to iterate review-and-fix with a specific reviewer. For skills prefer the skill-improver entry; for a branch prefer pr-improver.\"\nargument-hint: \"\u003CTARGET_PATH> --reviewer \u003Cagent-or-skill-name> --scope \u003Cglobs> [--max-rounds N]\"\nallowed-tools: Bash Glob Read TaskOutput TaskStop Workflow\n---\n\n# Code Improver\n\nImprove any code target by running `\u002Fcode-improver:improve` — a dynamic workflow that\nloops the named reviewer and a fixer subagent until a review reports zero critical\u002Fmajor\nfindings, then strips its own residue. The loop, its ledger, and its guards live in the\nworkflow; this skill collects the three inputs the generic entry requires and relays the\noutcome.\n\n## Starting the loop\n\nThe user provided: `$ARGUMENTS` (if empty, take the details from the conversation).\n\n### 1. Collect the three required inputs — no guessing\n\n1. **Target**: the absolute path to the directory under improvement. Resolve relative\n   paths against the working directory; verify the directory exists.\n2. **Reviewer**: the installed skill or agent that performs every review. The user must\n   name it — there is no default and no bundled reviewer. Determine the kind:\n   - a namespaced agent (e.g. `plugin-dev:skill-reviewer`) → `\"kind\": \"agent\"`\n   - an installed skill (e.g. `pr-review-toolkit:review-pr`) → `\"kind\": \"skill\"`\n   If the name could be either, check the session's skill listing; if still ambiguous,\n   ask the user. If no reviewer was named, ask — do not pick one.\n3. **Scope**: repo-relative globs the loop may touch. The generic entry requires it\n   explicitly; if the user did not give one, propose the target directory\n   (`\u003Crepo-relative-target>\u002F**`) and confirm before launching.\n\n### 2. Resolve the loop script\n\nThe loop is the dynamic workflow `workflows\u002Fimprove.js` in this plugin. Launch it by\npath: `scriptPath` takes a resolved absolute path, and the Workflow tool's `name` resolves\nbuilt-in and project workflows, so a marketplace-installed one may not answer to\n`code-improver:improve`. Try in order, first hit wins — the home directories come before\n`.` so an installed copy beats a checkout of this marketplace:\n\n1. `Bash: ls -d -- \"${CLAUDE_PLUGIN_ROOT}\u002Fworkflows\u002Fimprove.js\"`\n2. `Bash: ls -d -- \"${CODEX_PLUGIN_ROOT}\u002Fworkflows\u002Fimprove.js\"` (if that variable is set instead)\n3. `Bash: find ~\u002F.claude ~\u002F.codex . -maxdepth 7 -path '*\u002Fcode-improver\u002Fworkflows\u002Fimprove.js' -print -quit 2>\u002Fdev\u002Fnull`\n\nUse the path exactly as printed. Its plugin directory — the path with\n`\u002Fworkflows\u002Fimprove.js` removed — is `pluginRoot`. If all three come back empty, try\n`{name: \"code-improver:improve\"}` once; if that is unavailable too, stop and say the loop\ncould not be located. Do not assemble a path by hand and do not improvise the loop.\n\n### 3. Invoke the workflow\n\nRun it with the Workflow tool, `{scriptPath: \"\u003Cthe path from step 2>\", args: {...}}`:\n\n```json\n{\n  \"target\": \"\u003Cabsolute target path>\",\n  \"reviewer\": { \"kind\": \"agent|skill\", \"name\": \"\u003Cnamespaced-name>\", \"notes\": \"\u003Cwhat the reviewer should know about the target>\" },\n  \"scope\": [\"\u003Crepo-relative-glob>\u002F**\"],\n  \"pluginRoot\": \"\u003Cthe plugin directory from step 2>\",\n  \"maxRounds\": 5\n}\n```\n\n- `maxRounds` only if the user asked for a different cap.\n- `pluginRoot` lets the run find its metrics collector; omit the key only if step 2 fell\n  through to the workflow name — the workflow then searches for itself.\n- `finalize` (`{\"version_bump\": bool, \"narration_strip\": bool, \"docs_pass\": bool}`) only\n  to override the defaults: version bump when the target sits inside a plugin, narration\n  strip and docs pass always.\n- `decision` only on continuation (below).\n\nThe workflow runs in the background and needs no babysitting: it reviews, fixes,\nre-reviews, checks scope after every fix round, and can only complete on a clean review.\nIt never commits; all changes stay in the working tree.\n\n**If the Workflow tool is unavailable or denied, stop and say so.** Do not improvise the\nloop inline with direct edits — the ledger, scope guard, and escalation guarantees live\nin the workflow, and an inline imitation has none of them.\n\n**If the result is `halted: \"reviewer-unavailable\"`, relay it and stop.** The named\nreviewer is not installed in this session; tell the user which plugin provides it and\nre-run after installing. Do not review the target yourself.\n\n**Do not end your turn while the loop is running.** The Workflow tool returns a task id\nimmediately; the result comes later. In an interactive session the completion\nnotification re-invokes you — wait for it. In a non-interactive run (scripted, CI, eval)\nthere is no later turn: stopping abandons the loop mid-round, so after launching, poll\nthe task (TaskOutput with the returned task id, or sleep-and-recheck) until it completes,\nthen relay the result. A session that answers \"the loop is running, I'll report later\"\nhas lost the run.\n\n## Relaying the result\n\nThe workflow returns a structured result. Report it honestly — the distinctions matter:\n\n- **`converged: true`** — the last action was a review with zero critical\u002Fmajor findings.\n  Report rounds used, remaining minor findings (`open_minor_count`), and the artifact\n  paths (`ledger_path`, `metrics`).\n- **`capped: true`** — the fix budget ran out and the FINAL review still found blocking\n  issues. Say plainly: **capped, NOT converged**, and list `open_blocking`. Do not\n  present this as success.\n- **`escalation`** — the loop detected it was not converging (recurring findings,\n  non-decreasing counts, or a fix relocating a problem). Relay the escalation message and\n  finding ids to the user: this needs a design decision, not more rounds.\n- **`halted`** — a guard fired (scope violation, unregistered new files, a dead or\n  unavailable reviewer, or a finalize pass whose own edits failed the check that follows\n  it). Relay the paths in `violations`\u002F`new_untracked_files`, the sites in\n  `finalize_regressions`, and the notes.\n- **`notes`** always travel with the result — surface them; they include loud warnings\n  such as \"a git repository was initialized\".\n\n## Continuing after an escalation\n\nThe loop stops on escalation by design. When the user decides, start a fresh run with\nthe same `target` and `reviewer` plus:\n\n```json\n{ \"decision\": \"\u003Cthe user's ruling, verbatim>\" }\n```\n\nThe new run reloads the on-disk ledger, so every finding, rejection, and verdict carries\nover — rounds restart, re-derivation does not.\n\nTo stop a running loop, stop the workflow task (TaskStop); the ledger on disk is current\nto the last round and a re-run resumes from it.\n\n## When NOT to use\n\n- **A Claude Code skill**: use the `skill-improver` entry — it wires the right reviewer\n- **A branch \u002F pull request**: use the `pr-improver` entry — it derives scope from the diff\n- **One-time review**: dispatch the reviewer directly; the loop's value is iteration\n- **Quick single fixes**: edit the file directly\n","# Code Improver\n\nImprove any code target by running `\u002Fcode-improver:improve` — a dynamic workflow that\nloops the named reviewer and a fixer subagent until a review reports zero critical\u002Fmajor\nfindings, then strips its own residue. The loop, its ledger, and its guards live in the\nworkflow; this skill collects the three inputs the generic entry requires and relays the\noutcome.\n\n## Starting the loop\n\nThe user provided: `$ARGUMENTS` (if empty, take the details from the conversation).\n\n### 1. Collect the three required inputs — no guessing\n\n1. **Target**: the absolute path to the directory under improvement. Resolve relative\n   paths against the working directory; verify the directory exists.\n2. **Reviewer**: the installed skill or agent that performs every review. The user must\n   name it — there is no default and no bundled reviewer. Determine the kind:\n   - a namespaced agent (e.g. `plugin-dev:skill-reviewer`) → `\"kind\": \"agent\"`\n   - an installed skill (e.g. `pr-review-toolkit:review-pr`) → `\"kind\": \"skill\"`\n   If the name could be either, check the session's skill listing; if still ambiguous,\n   ask the user. If no reviewer was named, ask — do not pick one.\n3. **Scope**: repo-relative globs the loop may touch. The generic entry requires it\n   explicitly; if the user did not give one, propose the target directory\n   (`\u003Crepo-relative-target>\u002F**`) and confirm before launching.\n\n### 2. Resolve the loop script\n\nThe loop is the dynamic workflow `workflows\u002Fimprove.js` in this plugin. Launch it by\npath: `scriptPath` takes a resolved absolute path, and the Workflow tool's `name` resolves\nbuilt-in and project workflows, so a marketplace-installed one may not answer to\n`code-improver:improve`. Try in order, first hit wins — the home directories come before\n`.` so an installed copy beats a checkout of this marketplace:\n\n1. `Bash: ls -d -- \"${CLAUDE_PLUGIN_ROOT}\u002Fworkflows\u002Fimprove.js\"`\n2. `Bash: ls -d -- \"${CODEX_PLUGIN_ROOT}\u002Fworkflows\u002Fimprove.js\"` (if that variable is set instead)\n3. `Bash: find ~\u002F.claude ~\u002F.codex . -maxdepth 7 -path '*\u002Fcode-improver\u002Fworkflows\u002Fimprove.js' -print -quit 2>\u002Fdev\u002Fnull`\n\nUse the path exactly as printed. Its plugin directory — the path with\n`\u002Fworkflows\u002Fimprove.js` removed — is `pluginRoot`. If all three come back empty, try\n`{name: \"code-improver:improve\"}` once; if that is unavailable too, stop and say the loop\ncould not be located. Do not assemble a path by hand and do not improvise the loop.\n\n### 3. Invoke the workflow\n\nRun it with the Workflow tool, `{scriptPath: \"\u003Cthe path from step 2>\", args: {...}}`:\n\n```json\n{\n  \"target\": \"\u003Cabsolute target path>\",\n  \"reviewer\": { \"kind\": \"agent|skill\", \"name\": \"\u003Cnamespaced-name>\", \"notes\": \"\u003Cwhat the reviewer should know about the target>\" },\n  \"scope\": [\"\u003Crepo-relative-glob>\u002F**\"],\n  \"pluginRoot\": \"\u003Cthe plugin directory from step 2>\",\n  \"maxRounds\": 5\n}\n```\n\n- `maxRounds` only if the user asked for a different cap.\n- `pluginRoot` lets the run find its metrics collector; omit the key only if step 2 fell\n  through to the workflow name — the workflow then searches for itself.\n- `finalize` (`{\"version_bump\": bool, \"narration_strip\": bool, \"docs_pass\": bool}`) only\n  to override the defaults: version bump when the target sits inside a plugin, narration\n  strip and docs pass always.\n- `decision` only on continuation (below).\n\nThe workflow runs in the background and needs no babysitting: it reviews, fixes,\nre-reviews, checks scope after every fix round, and can only complete on a clean review.\nIt never commits; all changes stay in the working tree.\n\n**If the Workflow tool is unavailable or denied, stop and say so.** Do not improvise the\nloop inline with direct edits — the ledger, scope guard, and escalation guarantees live\nin the workflow, and an inline imitation has none of them.\n\n**If the result is `halted: \"reviewer-unavailable\"`, relay it and stop.** The named\nreviewer is not installed in this session; tell the user which plugin provides it and\nre-run after installing. Do not review the target yourself.\n\n**Do not end your turn while the loop is running.** The Workflow tool returns a task id\nimmediately; the result comes later. In an interactive session the completion\nnotification re-invokes you — wait for it. In a non-interactive run (scripted, CI, eval)\nthere is no later turn: stopping abandons the loop mid-round, so after launching, poll\nthe task (TaskOutput with the returned task id, or sleep-and-recheck) until it completes,\nthen relay the result. A session that answers \"the loop is running, I'll report later\"\nhas lost the run.\n\n## Relaying the result\n\nThe workflow returns a structured result. Report it honestly — the distinctions matter:\n\n- **`converged: true`** — the last action was a review with zero critical\u002Fmajor findings.\n  Report rounds used, remaining minor findings (`open_minor_count`), and the artifact\n  paths (`ledger_path`, `metrics`).\n- **`capped: true`** — the fix budget ran out and the FINAL review still found blocking\n  issues. Say plainly: **capped, NOT converged**, and list `open_blocking`. Do not\n  present this as success.\n- **`escalation`** — the loop detected it was not converging (recurring findings,\n  non-decreasing counts, or a fix relocating a problem). Relay the escalation message and\n  finding ids to the user: this needs a design decision, not more rounds.\n- **`halted`** — a guard fired (scope violation, unregistered new files, a dead or\n  unavailable reviewer, or a finalize pass whose own edits failed the check that follows\n  it). Relay the paths in `violations`\u002F`new_untracked_files`, the sites in\n  `finalize_regressions`, and the notes.\n- **`notes`** always travel with the result — surface them; they include loud warnings\n  such as \"a git repository was initialized\".\n\n## Continuing after an escalation\n\nThe loop stops on escalation by design. When the user decides, start a fresh run with\nthe same `target` and `reviewer` plus:\n\n```json\n{ \"decision\": \"\u003Cthe user's ruling, verbatim>\" }\n```\n\nThe new run reloads the on-disk ledger, so every finding, rejection, and verdict carries\nover — rounds restart, re-derivation does not.\n\nTo stop a running loop, stop the workflow task (TaskStop); the ledger on disk is current\nto the last round and a re-run resumes from it.\n\n## When NOT to use\n\n- **A Claude Code skill**: use the `skill-improver` entry — it wires the right reviewer\n- **A branch \u002F pull request**: use the `pr-improver` entry — it derives scope from the diff\n- **One-time review**: dispatch the reviewer directly; the loop's value is iteration\n- **Quick single fixes**: edit the file directly\n",{"name":5,"description":7,"allowed-tools":23,"argument-hint":24},"Bash Glob Read TaskOutput TaskStop Workflow","\u003CTARGET_PATH> --reviewer \u003Cagent-or-skill-name> --scope \u003Cglobs> [--max-rounds N]","main","321ccfe628eca0d314b0ee4eaffcdd8a05639aaf","plugins\u002Fcode-improver\u002Fskills\u002Fcode-improver",0,"https:\u002F\u002Fgithub.com\u002Fbongkage\u002Ftrailofbits-skills","https:\u002F\u002Fgithub.com\u002Fbongkage","2026-11-06 12:15:06",{"source":8,"skillName":5,"commands":33,"prompt":43,"zipAvailable":44},[34,37,40],{"tool":35,"command":36},"npx","npx skills add bongkage\u002Ftrailofbits-skills --skill code-improver",{"tool":38,"command":39},"bunx","bunx skills add bongkage\u002Ftrailofbits-skills --skill code-improver",{"tool":41,"command":42},"pnpm","pnpm dlx skills add bongkage\u002Ftrailofbits-skills --skill code-improver","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fbongkage\u002Ftrailofbits-skills\u002Fblob\u002Fmain\u002Fplugins\u002Fcode-improver\u002Fskills\u002Fcode-improver\u002FSKILL.md",true,[],[47,54,61,68,75,82],{"skillNo":48,"name":49,"nameZh":6,"description":50,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":51,"categories":52,"keywords":53,"publishedAt":18},"01M7CHJAESZRWKJXFQG55BMNAA","zeroize-audit","Detects missing zeroization of sensitive data in source code and identifies zeroization removed by compiler optimizations, with assembly-level analysis, and control-flow verification. Use for auditing C\u002FC++\u002FRust code handling secrets, keys, passwords, or other sensitive data.","https:\u002F\u002Fgithub.com\u002Fbongkage\u002Ftrailofbits-skills\u002Fblob\u002Fmain\u002Fplugins\u002Fzeroize-audit\u002Fskills\u002Fzeroize-audit\u002FSKILL.md",[],[],{"skillNo":55,"name":56,"nameZh":6,"description":57,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":58,"categories":59,"keywords":60,"publishedAt":18},"01M7CHJ543G42A9GQC3PJE4F4V","yara-rule-authoring","Guides authoring of high-quality YARA-X detection rules for malware identification. Use when writing, reviewing, or optimizing YARA rules. Covers naming conventions, string selection, performance optimization, migration from legacy YARA, and false positive reduction. Triggers on: YARA, YARA-X, malware detection, threat hunting, IOC, signature, crx module, dex module.","https:\u002F\u002Fgithub.com\u002Fbongkage\u002Ftrailofbits-skills\u002Fblob\u002Fmain\u002Fplugins\u002Fyara-authoring\u002Fskills\u002Fyara-rule-authoring\u002FSKILL.md",[],[],{"skillNo":62,"name":63,"nameZh":6,"description":64,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":65,"categories":66,"keywords":67,"publishedAt":18},"01M7CHHYVFS1AKDTH3YDS1Y4K4","writing-lean-proofs","Writes and reviews structured Lean 4 proofs and designs Lean libraries following Mathlib conventions. Use when proving theorems in Lean, formalizing mathematics or specifications in Lean 4, defining new types or definitions in a Lean library, reviewing Lean proofs for readability and maintainability, refactoring long tactic proofs into lemmas, filling in sorry placeholders in a Lean development, setting up CI or linters for a Lean project, diagnosing slow proofs or maxHeartbeats timeouts, or writing custom tactics, macros, or linters.","https:\u002F\u002Fgithub.com\u002Fbongkage\u002Ftrailofbits-skills\u002Fblob\u002Fmain\u002Fplugins\u002Fwriting-lean-proofs\u002Fskills\u002Fwriting-lean-proofs\u002FSKILL.md",[],[],{"skillNo":69,"name":70,"nameZh":6,"description":71,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":72,"categories":73,"keywords":74,"publishedAt":18},"01M7CHHQXA39M4965MV2RMM5YS","vulnerability-triage-brocards","This skill should be used when the user asks to \"triage a vulnerability report\", \"assess a CVE\", \"evaluate a bug bounty submission\", \"decide if a finding is valid\", \"review a security finding\", \"dismiss a vulnerability\", \"should we fix this CVE\", \"prioritize a vulnerability report\", or needs to determine whether an incoming vulnerability report warrants investigation. Applies 7 brocards (rules of thumb) to systematically accept, dismiss, or request more information on vulnerability reports, or needs to filter raw findings from agentic vulnerability discovery pipelines before human review.","https:\u002F\u002Fgithub.com\u002Fbongkage\u002Ftrailofbits-skills\u002Fblob\u002Fmain\u002Fplugins\u002Fvulnerability-triage-brocards\u002Fskills\u002Fvulnerability-triage-brocards\u002FSKILL.md",[],[],{"skillNo":76,"name":77,"nameZh":6,"description":78,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":79,"categories":80,"keywords":81,"publishedAt":18},"01M7CHHJVQC2ZG2M310GXT0T42","variant-analysis","Hunts for the other instances of a bug already found — the variants of one root cause across a codebase. Use immediately after a vulnerability, logic bug, or bad pattern turns up in a specific file and the question becomes where else it occurs, including the bare conversational form (\"are there others like this?\", \"is this the same bug?\"). Also for generalizing one known instance into a CodeQL or Semgrep query for its whole pattern family, and for triaging a set of look-alike candidates against a known root cause. Not for initial discovery with no bug in hand.","https:\u002F\u002Fgithub.com\u002Fbongkage\u002Ftrailofbits-skills\u002Fblob\u002Fmain\u002Fplugins\u002Fvariant-analysis\u002Fskills\u002Fvariant-analysis\u002FSKILL.md",[],[],{"skillNo":83,"name":84,"nameZh":6,"description":85,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":86,"categories":87,"keywords":88,"publishedAt":18},"01M7CHHC1GM5GH585ZWGP7D03P","vector-forge","Mutation-driven test vector generation. Finds implementations of a cryptographic algorithm or protocol, runs mutation testing to identify escaped mutants, then generates new test vectors that deliberately exercise the uncovered code paths. Compares before\u002Fafter mutation kill rates to prove vector effectiveness. Use when generating cryptographic test vectors, measuring Wycheproof coverage gaps, finding escaped mutants via mutation testing, creating cross-implementation test suites, or improving test vector coverage for crypto primitives.","https:\u002F\u002Fgithub.com\u002Fbongkage\u002Ftrailofbits-skills\u002Fblob\u002Fmain\u002Fplugins\u002Ftrailmark\u002Fskills\u002Fvector-forge\u002FSKILL.md",[],[]]