A Claude Code skill bundle for bug hunting and external red-team work - 82 skills, 15 slash commands, 681 disclosed-report patterns curated across 24 core vulnerability classes, plus enterprise identity + infrastructure attack matrices.
Python ai-securityanthropicapplication-securitybug-bountybugbountybugcrowdclaudeclaude-codeclaude-skillsethical-hackinghackeroneoffensive-securitypentestingred-teamsecurity-toolsweb-security
ci+community: add skill-linter, issue/PR templates, COC, CHANGELOG, FUNDING
Sachin Sharmabb9d93587b68
diff --git a/.github/CODEOWNERS b/.github/CODEOWNERSnew file mode 100644index 0000000..45f6fce--- /dev/null+++ b/.github/CODEOWNERS@@ -0,0 +1,8 @@+# Default owner for everything — gets auto-requested on every PR.+* @elementalsouls++# Safety-critical files: changes here should always be maintainer-reviewed.+/scripts/lint_skills.py @elementalsouls+/scripts/.identifier-denylist.sha256 @elementalsouls+/.github/ @elementalsouls+/SECURITY.md @elementalsoulsdiff --git a/.github/FUNDING.yml b/.github/FUNDING.ymlnew file mode 100644index 0000000..be4ccf0--- /dev/null+++ b/.github/FUNDING.yml@@ -0,0 +1,5 @@+# GitHub Sponsors / funding links shown on the repo's "Sponsor" button.+# Sponsor: Atlas Cloud — full-modal AI inference platform (project sponsor).+github: [elementalsouls]+custom:+ - "https://www.atlascloud.ai/console/coding-plan"diff --git a/.github/ISSUE_TEMPLATE/bug_report.yml b/.github/ISSUE_TEMPLATE/bug_report.ymlnew file mode 100644index 0000000..f9e5803--- /dev/null+++ b/.github/ISSUE_TEMPLATE/bug_report.yml@@ -0,0 +1,45 @@+name: 🐞 Bug / incorrect skill behavior+description: A skill misfires, fails to trigger, gives wrong guidance, or a script errors.+title: "[bug] <short summary>"+labels: ["bug"]+body:+ - type: markdown+ attributes:+ value: |+ Thanks for reporting. Please **do not paste real client/engagement data, credentials, or target identifiers** — anonymize everything (see CONTRIBUTING.md).+ - type: input+ id: skill+ attributes:+ label: Which skill / command?+ placeholder: "e.g. hunt-ssrf, /hunt, scripts/cbh.py"+ validations:+ required: true+ - type: textarea+ id: expected+ attributes:+ label: What did you expect to happen?+ validations:+ required: true+ - type: textarea+ id: actual+ attributes:+ label: What actually happened?+ description: Include the prompt you used and what Claude did (or didn't do). Redact any sensitive data.+ validations:+ required: true+ - type: dropdown+ id: trigger+ attributes:+ label: Did the skill auto-trigger?+ options:+ - "Yes, it loaded but behaved wrong"+ - "No, it never triggered when it should have"+ - "It triggered when it should NOT have"+ - "N/A"+ validations:+ required: true+ - type: input+ id: env+ attributes:+ label: Environment+ placeholder: "Claude Code version, OS, install method (symlink/copy)"diff --git a/.github/ISSUE_TEMPLATE/config.yml b/.github/ISSUE_TEMPLATE/config.ymlnew file mode 100644index 0000000..d467496--- /dev/null+++ b/.github/ISSUE_TEMPLATE/config.yml@@ -0,0 +1,8 @@+blank_issues_enabled: false+contact_links:+ - name: 💬 Discussions (questions, ideas, show-and-tell)+ url: https://github.com/elementalsouls/Claude-BugHunter/discussions+ about: Ask how to use a skill, request a new one, or share a hunt write-up. Not for bugs.+ - name: 🔒 Security / responsible disclosure+ url: https://github.com/elementalsouls/Claude-BugHunter/blob/main/SECURITY.md+ about: Report a vulnerability in the tooling itself, or a misuse concern. Do not open a public issue.diff --git a/.github/ISSUE_TEMPLATE/false_positive.yml b/.github/ISSUE_TEMPLATE/false_positive.ymlnew file mode 100644index 0000000..d0445b0--- /dev/null+++ b/.github/ISSUE_TEMPLATE/false_positive.yml@@ -0,0 +1,33 @@+name: 🎯 False-positive / bad guidance report+description: A skill led to a finding that turned out to be invalid, or gave guidance that breaks the false-positive discipline.+title: "[fp] <skill> — <what was wrong>"+labels: ["false-positive"]+body:+ - type: markdown+ attributes:+ value: |+ This project takes false-positive discipline seriously (OOB confirmation, unique markers, body-diff, server-policy-vs-state). Help us tighten the skills. **Anonymize all target data.**+ - type: input+ id: skill+ attributes:+ label: Which skill produced the bad guidance?+ validations:+ required: true+ - type: textarea+ id: claim+ attributes:+ label: What did the skill claim / suggest?+ validations:+ required: true+ - type: textarea+ id: reality+ attributes:+ label: Why was it actually invalid?+ description: e.g. reflected value was server policy not attacker-controlled state; no OOB confirmation; response diff was noise.+ validations:+ required: true+ - type: textarea+ id: fix+ attributes:+ label: Suggested guardrail+ description: What validation step or wording would have prevented it?diff --git a/.github/ISSUE_TEMPLATE/new_skill.yml b/.github/ISSUE_TEMPLATE/new_skill.ymlnew file mode 100644index 0000000..231373a--- /dev/null+++ b/.github/ISSUE_TEMPLATE/new_skill.yml@@ -0,0 +1,44 @@+name: ✨ New skill proposal+description: Propose a new hunt-* or capability skill. Open this BEFORE writing the SKILL.md (per CONTRIBUTING.md).+title: "[skill] hunt-<name> — <one-line purpose>"+labels: ["skill-proposal"]+body:+ - type: markdown+ attributes:+ value: |+ Per CONTRIBUTING.md, propose new skills here first so we can confirm scope and avoid overlap before you write the full SKILL.md.+ - type: input+ id: name+ attributes:+ label: Proposed skill name+ description: lowercase-hyphen only, e.g. hunt-fintech-graphql+ validations:+ required: true+ - type: textarea+ id: purpose+ attributes:+ label: What gap does it fill?+ description: What can't be done well with the existing 51 skills? Be specific.+ validations:+ required: true+ - type: textarea+ id: overlap+ attributes:+ label: Overlap check+ description: Which existing skills are closest, and why is this distinct enough to stand alone?+ validations:+ required: true+ - type: textarea+ id: sources+ attributes:+ label: Source material+ description: Disclosed reports, CVEs, or research this would be built from (links preferred — the house style cites named reports).+ - type: checkboxes+ id: scope+ attributes:+ label: Scope confirmation+ options:+ - label: This is external-attack-surface focused (not internal AD / C2 / post-exploit — those are deliberately out of scope).+ required: true+ - label: It contains no real client/engagement identifiers, credentials, or target-specific exploits — all class-based / anonymized.+ required: truediff --git a/.github/PULL_REQUEST_TEMPLATE.md b/.github/PULL_REQUEST_TEMPLATE.mdnew file mode 100644index 0000000..85e652a--- /dev/null+++ b/.github/PULL_REQUEST_TEMPLATE.md@@ -0,0 +1,26 @@+<!-- Thanks for contributing to claude-bughunter. Keep PRs focused. -->++## What does this change?++<!-- One or two sentences. Link the proposal issue for new skills (#NN). -->++## Type++- [ ] New skill+- [ ] Improvement to an existing skill+- [ ] Docs / examples+- [ ] Tooling / scripts / CI+- [ ] Other:++## Checklist++- [ ] **No client/engagement data** — no real target names, account UIDs, endpoints, bounty amounts, credentials, or PII. All examples are class-based / anonymized.+- [ ] `python3 scripts/lint_skills.py` passes locally (frontmatter, name, description length, identifier/secret scan).+- [ ] For a **new skill**: a proposal issue was opened and agreed first; `name` is lowercase-hyphen; description ≤ 1024 chars and weaves in trigger keywords; body ≤ ~500 lines.+- [ ] For a new skill: mentioned in `README.md` + `USAGE.md` decision tree, with one worked trigger example.+- [ ] Sources cited for anything adapted from disclosed reports / community work.+- [ ] Cross-references complementary skills ("see also …") where there's topical overlap.++## Anything reviewers should know?++<!-- Overlap with existing skills, follow-ups, open questions. -->diff --git a/.github/workflows/skill-lint.yml b/.github/workflows/skill-lint.ymlnew file mode 100644index 0000000..f2f13b1--- /dev/null+++ b/.github/workflows/skill-lint.yml@@ -0,0 +1,41 @@+name: skill-lint++# Quality + safety gate for every skill change.+# - Validates SKILL.md structure against CONTRIBUTING.md rules+# - Scans for client/engagement identifiers (hashed denylist) and leaked secrets+# Runs on PRs and pushes to main. No external dependencies — stdlib Python only.++on:+ pull_request:+ paths:+ - "skills/**"+ - "scripts/lint_skills.py"+ - "scripts/.identifier-denylist.sha256"+ - ".github/workflows/skill-lint.yml"+ push:+ branches: [main]+ paths:+ - "skills/**"+ - "scripts/lint_skills.py"+ workflow_dispatch:++permissions:+ contents: read++jobs:+ lint:+ name: Lint skills+ runs-on: ubuntu-latest+ steps:+ - uses: actions/checkout@v4++ - uses: actions/setup-python@v5+ with:+ python-version: "3.12"++ - name: Validate skills (structure + identifier/secret scan)+ run: python3 scripts/lint_skills.py++ - name: Summary+ if: always()+ run: echo "Skill lint finished — see annotations above for any errors/warnings."diff --git a/.gitignore b/.gitignoreindex 07d794e..844c547 100644--- a/.gitignore+++ b/.gitignore@@ -58,3 +58,6 @@ node_modules/# Local cbh recon outputs (when run from repo root)recon/++# Maintainer-only plaintext client-identifier denylist (never commit names)+scripts/.identifier-denylist.localdiff --git a/CHANGELOG.md b/CHANGELOG.mdnew file mode 100644index 0000000..8f61dec--- /dev/null+++ b/CHANGELOG.md@@ -0,0 +1,42 @@+# Changelog++All notable changes to this project are documented here.+Format follows [Keep a Changelog](https://keepachangelog.com/en/1.1.0/);+versioning is loosely [SemVer](https://semver.org/) at the bundle level.++## [Unreleased]++### Added+- **CI skill-linter** (`scripts/lint_skills.py` + `.github/workflows/skill-lint.yml`) —+ validates every `SKILL.md` for frontmatter, `name` format, description length, and+ body length per `CONTRIBUTING.md`, and scans for leaked secrets and client/engagement+ identifiers. The identifier denylist is stored as SHA-256 hashes+ (`scripts/.identifier-denylist.sha256`) so client names never enter the repo in plaintext.+- **Community infrastructure** — issue templates (bug, new-skill proposal, false-positive),+ PR template, `CODEOWNERS`, `FUNDING.yml`, `CODE_OF_CONDUCT.md`, and this `CHANGELOG.md`.+- **Sponsor** — Atlas Cloud added to the README (theme-adaptive logo).++### Changed+- `.gitignore` now excludes the maintainer-only plaintext denylist override+ (`scripts/.identifier-denylist.local`).++<!-- When PR #7 (20 new hunt-* skills) merges, move its entry up to the next release. -->++## [2.0] - 2026-05-25++### Added+- Report-curation pass: 574 → **681 disclosed-report patterns** across 24 vulnerability classes.+- 5 previously-missing attack surfaces covered; 0 zero-report skills remaining.+- 29 A-to-B chain examples and `ENGAGEMENTS.md` scaffolding.+- Enterprise platform attack matrices (M365/Entra, Okta, SharePoint, vCenter, SSL-VPN, APK, supply-chain).++### Changed+- Top-3 trigger-match concentration rebalanced (81.2% → 68.4%) for better skill routing.++## [1.x]++- Initial public release: 51 skills + 15 slash commands, vendored foundation from+ `shuvonsec/claude-bug-bounty`, Burp MCP integration, recon pipeline.++[Unreleased]: https://github.com/elementalsouls/Claude-BugHunter/compare/v2.0...HEAD+[2.0]: https://github.com/elementalsouls/Claude-BugHunter/releases/tag/v2.0diff --git a/CODE_OF_CONDUCT.md b/CODE_OF_CONDUCT.mdnew file mode 100644index 0000000..fe6d71d--- /dev/null+++ b/CODE_OF_CONDUCT.md@@ -0,0 +1,63 @@+# Code of Conduct++## Our pledge++We — contributors and maintainers — pledge to make participation in this project+a harassment-free experience for everyone, regardless of age, body size, visible+or invisible disability, ethnicity, sex characteristics, gender identity and+expression, level of experience, education, socio-economic status, nationality,+personal appearance, race, religion, or sexual identity and orientation.++We pledge to act and interact in ways that contribute to an open, welcoming,+diverse, inclusive, and healthy community.++## Our standards++Examples of behavior that contributes to a positive environment:++- Being respectful of differing opinions, viewpoints, and experiences+- Giving and gracefully accepting constructive feedback on skills and PRs+- Accepting responsibility, apologizing to those affected by our mistakes, and+ learning from the experience+- Focusing on what is best for the community and the people who rely on this tooling++Examples of unacceptable behavior:++- Trolling, insulting or derogatory comments, and personal or political attacks+- Public or private harassment+- Publishing others' private information without explicit permission+- Sexualized language or imagery, or unwelcome sexual attention+- Other conduct which could reasonably be considered inappropriate in a+ professional setting++## Security-tooling responsible-use clause++This project ships offensive-security tradecraft. In addition to the above:++- **Do not** use issues, discussions, or PRs to solicit, coordinate, or brag about+ testing against systems you are not authorized to assess. See `SECURITY.md` for+ the authorized-use posture.+- **Do not** post real client/engagement identifiers, credentials, victim PII, or+ target-specific exploit details. Anonymize everything (see `CONTRIBUTING.md`).+- Discussion of techniques is welcome; using this community to facilitate+ unauthorized access or harm is not, and is grounds for an immediate ban.++## Enforcement++Instances of abusive, harassing, or otherwise unacceptable behavior — or misuse of+the community as described above — may be reported to the maintainer at+**[email protected]**. All complaints will be reviewed and investigated+promptly and fairly. The maintainer is obligated to respect the privacy and+security of the reporter.++Maintainers who do not follow or enforce this Code of Conduct in good faith may+face temporary or permanent repercussions as determined by the project's leadership.++## Attribution++This Code of Conduct is adapted from the [Contributor Covenant][homepage],+version 2.1, available at+<https://www.contributor-covenant.org/version/2/1/code_of_conduct.html>,+with an added responsible-use clause specific to offensive-security tooling.++[homepage]: https://www.contributor-covenant.orgdiff --git a/scripts/.identifier-denylist.sha256 b/scripts/.identifier-denylist.sha256new file mode 100644index 0000000..3c16ea7--- /dev/null+++ b/scripts/.identifier-denylist.sha256@@ -0,0 +1,22 @@+# Client-identifier denylist — SHA-256 hashes, NOT plaintext.+#+# Why hashes? The whole point of this gate is to keep real client/engagement+# identifiers OUT of the public repo. Committing the names in plaintext (even in+# a denylist) would itself leak them. So we store SHA-256 of each normalized+# identifier instead. The linter (scripts/lint_skills.py) normalizes every+# 1- and 2-word shingle of each SKILL.md the same way and compares hashes.+#+# Normalization: lowercase, collapse internal whitespace to single spaces.+#+# To add a NEW identifier without leaking it, either:+# (a) add its SHA-256 here: printf '%s' "the name" | tr '[:upper:]' '[:lower:]' | shasum -a 256+# (b) drop it in scripts/.identifier-denylist.local (one per line, gitignored) —+# the linter hashes that file at runtime so you never commit the plaintext.+#+# One hash per line. Blank lines and #-comments ignored.+a672b3be0c15a2af015886aef804d485704c37a56cfa8a564c973ccd944a2ee1+6d99af9a927fd75a379a2d085c772acd6fb4a81cd7a3c75aaffa7529536d0e1a+3b39f05d6785c7b4ac3e3e08dfc2abbc07bdde4a01c58fdf944d326afd4af3ab+80537d65d479a2f986a78e859cc2bcaaf6a62f6df7a6ba2c5dd192cfcb2ff790+ae95921ebc9d9f3e72d883f217ff023fc5555d111ca0bf3b18c5b8cd0c5d7580+e7dcc519de2734ed30de133910fd3289c9afc89aef272f7239d17fac804b24a0diff --git a/scripts/lint_skills.py b/scripts/lint_skills.pynew file mode 100644index 0000000..2b9e93b--- /dev/null+++ b/scripts/lint_skills.py@@ -0,0 +1,205 @@+#!/usr/bin/env python3+"""+lint_skills.py — quality + safety gate for claude-bughunter skills.++Enforces the rules documented in CONTRIBUTING.md and the repo's hard rule that+NO real client/engagement identifiers ever land in the public tree.++Checks per skills/<name>/SKILL.md:+ STRUCTURE (errors)+ - frontmatter block present (between the first two `---` lines)+ - `name` present, matches ^[a-z0-9-]+$, and equals the directory name+ - `description` present and <= 1024 chars+ - body (everything after frontmatter) <= 500 lines+ SAFETY (errors)+ - client-identifier denylist: hashes every 1- and 2-word shingle of the file+ and compares against scripts/.identifier-denylist.sha256 (+ optional+ .identifier-denylist.local). Plaintext names never live in the repo.+ - real-secret scan: AWS keys, private-key blocks, JWTs, Slack/Google tokens.+ Documentation patterns (regexes, AWS EXAMPLE tokens) are allowlisted so the+ secret-catalog skills don't trip it.++Exit code 0 = clean, 1 = at least one error. Warnings never fail the build.+Stdlib only — no pip install needed in CI.++Usage:+ python3 scripts/lint_skills.py # lint all skills+ python3 scripts/lint_skills.py skills/hunt-xss # lint specific dirs+"""+import hashlib+import os+import re+import sys++REPO = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))+SKILLS_DIR = os.path.join(REPO, "skills")+SCRIPTS_DIR = os.path.join(REPO, "scripts")++NAME_RE = re.compile(r"^[a-z0-9-]+$")+MAX_DESC = 1024+MAX_BODY_LINES = 500++# --- real-secret patterns (kept tight to avoid flagging documented regexes) ---+SECRET_PATTERNS = [+ ("AWS access key", re.compile(r"\bAKIA[0-9A-Z]{16}\b")),+ ("Private key block", re.compile(r"-----BEGIN (?:RSA |EC |OPENSSH |DSA |PGP )?PRIVATE KEY-----")),+ ("JWT", re.compile(r"\beyJ[A-Za-z0-9_-]{8,}\.eyJ[A-Za-z0-9_-]{8,}\.[A-Za-z0-9_-]{8,}\b")),+ ("Slack token", re.compile(r"\bxox[baprs]-[0-9A-Za-z-]{10,}\b")),+ ("Google API key", re.compile(r"\bAIza[0-9A-Za-z_-]{35}\b")),+ ("GitHub PAT", re.compile(r"\bghp_[0-9A-Za-z]{36}\b")),+]+# Tokens that are public documentation/examples, not real secrets.+# `\.\.\.` covers placeholder blocks like "-----BEGIN PRIVATE KEY-----..." in docs.+SECRET_ALLOW = re.compile(+ r"AKIAIOSFODNN7EXAMPLE|EXAMPLE|wJalrXUtnFEMI|\[0-9A-Z\]|\{1[06]\}|<[^>]+>|\.\.\."+)++# Intentional kitchen-sink router/aggregator skills whose descriptions deliberately+# exceed the per-skill limit (they route to everything). Over-length is a warning,+# not an error, for these. Do NOT add new skills here — write focused descriptions.+DESC_LIMIT_GRANDFATHERED = {"bug-bounty", "bb-local-toolkit", "osint-methodology"}++WORD_RE = re.compile(r"[a-z0-9]+")+++def load_denylist():+ """Return a set of sha256 hex digests of banned identifiers."""+ hashes = set()+ sha_file = os.path.join(SCRIPTS_DIR, ".identifier-denylist.sha256")+ if os.path.exists(sha_file):+ with open(sha_file, encoding="utf-8") as fh:+ for line in fh:+ line = line.strip()+ if line and not line.startswith("#"):+ hashes.add(line.lower())+ # Optional gitignored plaintext override for maintainer convenience.+ local = os.path.join(SCRIPTS_DIR, ".identifier-denylist.local")+ if os.path.exists(local):+ with open(local, encoding="utf-8") as fh:+ for line in fh:+ name = " ".join(line.strip().lower().split())+ if name and not name.startswith("#"):+ hashes.add(hashlib.sha256(name.encode()).hexdigest())+ return hashes+++def shingles(text):+ """Yield normalized 1- and 2-word shingles from text."""+ words = WORD_RE.findall(text.lower())+ for i, w in enumerate(words):+ yield w+ if i + 1 < len(words):+ yield w + " " + words[i + 1]+++def split_frontmatter(raw):+ """Return (frontmatter_dict, body_text, error_or_None)."""+ if not raw.startswith("---"):+ return {}, raw, "no frontmatter block (file must start with '---')"+ parts = raw.split("\n")+ # find closing --- after line 0+ close = None+ for i in range(1, len(parts)):+ if parts[i].strip() == "---":+ close = i+ break+ if close is None:+ return {}, raw, "frontmatter opened with '---' but never closed"+ fm = {}+ for line in parts[1:close]:+ if not line.strip() or line.lstrip().startswith("#"):+ continue+ m = re.match(r"^([A-Za-z0-9_-]+):\s*(.*)$", line)+ if m:+ fm[m.group(1)] = m.group(2).strip()+ body = "\n".join(parts[close + 1:])+ return fm, body, None+++def lint_skill(skill_dir, denylist):+ errors, warnings = [], []+ name = os.path.basename(skill_dir.rstrip("/"))+ path = os.path.join(skill_dir, "SKILL.md")+ if not os.path.isfile(path):+ return [f"{name}: missing SKILL.md"], []+ with open(path, encoding="utf-8") as fh:+ raw = fh.read()++ fm, body, fm_err = split_frontmatter(raw)+ if fm_err:+ errors.append(f"{name}: {fm_err}")+ # name+ fn = fm.get("name", "")+ if not fn:+ errors.append(f"{name}: frontmatter missing `name`")+ else:+ if not NAME_RE.match(fn):+ errors.append(f"{name}: `name` '{fn}' must match ^[a-z0-9-]+$")+ if fn != name:+ errors.append(f"{name}: `name` '{fn}' != directory '{name}'")+ # description+ desc = fm.get("description", "")+ if not desc:+ errors.append(f"{name}: frontmatter missing `description`")+ elif len(desc) > MAX_DESC:+ msg = f"{name}: description {len(desc)} chars > {MAX_DESC} limit"+ (warnings if name in DESC_LIMIT_GRANDFATHERED else errors).append(msg)+ elif len(desc) < 40:+ warnings.append(f"{name}: description very short ({len(desc)} chars) — weak trigger surface")+ # body length+ body_lines = len(body.splitlines())+ if body_lines > MAX_BODY_LINES:+ warnings.append(f"{name}: body {body_lines} lines > {MAX_BODY_LINES} guideline (use references/ subfolder)")++ # client-identifier denylist+ if denylist:+ hit = set()+ for sh in shingles(raw):+ if hashlib.sha256(sh.encode()).hexdigest() in denylist:+ hit.add(sh)+ if hit:+ errors.append(+ f"{name}: CLIENT-IDENTIFIER MATCH — {len(hit)} banned shingle(s). "+ f"Remove client/engagement identifiers before committing."+ )++ # real-secret scan+ for lineno, line in enumerate(raw.splitlines(), 1):+ for label, pat in SECRET_PATTERNS:+ for m in pat.finditer(line):+ snippet = m.group(0)+ window = line[max(0, m.start() - 8): m.end() + 8]+ if SECRET_ALLOW.search(window) or SECRET_ALLOW.search(snippet):+ continue+ errors.append(f"{name}:{lineno}: possible {label} leaked — '{snippet[:24]}...'")+ return errors, warnings+++def main(argv):+ denylist = load_denylist()+ if argv:+ targets = [a if os.path.isabs(a) else os.path.join(REPO, a) for a in argv]+ else:+ targets = [os.path.join(SKILLS_DIR, d) for d in sorted(os.listdir(SKILLS_DIR))+ if os.path.isdir(os.path.join(SKILLS_DIR, d))]++ all_errors, all_warnings = [], []+ for t in targets:+ e, w = lint_skill(t, denylist)+ all_errors += e+ all_warnings += w++ for w in all_warnings:+ print(f"::warning:: {w}" if os.environ.get("GITHUB_ACTIONS") else f"WARN {w}")+ for e in all_errors:+ print(f"::error:: {e}" if os.environ.get("GITHUB_ACTIONS") else f"ERROR {e}")++ n = len(targets)+ print(f"\nLinted {n} skill(s): {len(all_errors)} error(s), {len(all_warnings)} warning(s).")+ if not denylist:+ print("NOTE: no client-identifier denylist loaded (scripts/.identifier-denylist.sha256 missing).")+ return 1 if all_errors else 0+++if __name__ == "__main__":+ sys.exit(main(sys.argv[1:]))