| Files | Last commit | Last update |
|---|---|---|
security: repo-wide client-identifier leak guard (fail-closed) + pre-commit hook (#65) The existing hashed-denylist scan lives in lint_skills.py and only covers skills/**. The identifier leak that motivated this guard landed in README.md and tests/ — files that linter never sees, and skill-lint.yml only triggers on skills/** paths anyway. It also fails OPEN: a missing denylist prints a NOTE and exits 0. This closes both holes: - scripts/scan_identifiers.py — standalone scanner reusing the same denylist file and 1-/2-word shingle-hash logic, but over EVERY tracked text file (git ls-files, minus binaries). Fails CLOSED: missing/empty denylist = exit 2. - .github/workflows/leak-guard.yml — runs it (and its self-test) on every push and PR with NO path filter. contents: read. - .githooks/pre-commit — same scan locally before push (opt-in via `git config core.hooksPath .githooks`). - scripts/test_scan_identifiers.py — stdlib self-check: catch, clean, case-insensitive, and fail-closed behaviors. - scripts/.identifier-denylist.sha256 — added the single-word hash to complement the existing two-word entry (both now covered). - CONTRIBUTING.md — documents the guard, the hook, and how to ban a new identifier without committing the plaintext. Verified: whole-repo scan 0 matches / exit 0; self-test passes; missing denylist exits 2; a planted identifier is blocked by both the scanner and the hook; lint_skills.py still 0 errors. (The guard even caught a client name I'd left in this scanner's own docstring — removed.) Scope note: identifier-only by design. Repo-wide secret scanning is deferred — it would false-positive on the synthetic AKIA test keys in docs/verification/, which need allowlist work first. Secrets stay covered by lint_skills.py (skills) and secret_scan.py. Co-authored-by: Sachin Sharma <elementalsoul@Sachins-MacBook-Pro.local> Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com> | 2 months ago | |
DX hardening, cbh packaging, and Claude-OSINT integration (#43) * fix(dx): harden install, package cbh as pipx-installable, fix broken commands Addresses 6 P1 defects from a DX review (solo bug-bounty hunter persona): - install: .gitattributes (eol=lf) + CRLF self-heal in install.sh; INSTALL uses bash form - cli: package cbh as a real pip/pipx console-script (pyproject.toml + cbh/ package + bundled cbh/data/skill_index.json fallback + scripts/gen_skill_index.py); scripts/cbh.py kept as a clone shim; recon writes to ./recon when installed - commands: add /scope backed by a new deterministic engine/scope.py CLI (was called mandatory in /hunt but missing); reframe /token-scan, /intel, /surface so the skill path is primary and missing helper scripts are clearly optional - docs: install capability matrix (plugin vs clone vs pipx); reconcile the authorization story in /hunt (don't ingest the SOW document != skip scope verification) Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * feat(skills): consolidate recon skills as canonical (integration Phase 1) BugHunter becomes the source of truth for the two recon skills shared with Claude-OSINT (which now re-exports them). - offensive-osint: port h1_reference.py (HackerOne hacktivity agent) + dashboard.py (local recon console) + assets/ font from Claude-OSINT (CRLF->LF); documented in references/helpers-and-automation.md sec 49-50 + SKILL.md router. references/ split kept; OSINT's recent change was dashboard-only so no SKILL.md content delta. - osint-methodology: keep BugHunter's expanded version (1703 vs 455 lines, supersedes OSINT's); bump version 2.1 -> 2.3 to pass OSINT's 2.2. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * fix(dx): correct CRLF docs + scope.py footguns (post-verification) A 20-agent adversarial verification found the install.sh CRLF self-heal was non-functional (bash parse-aborts on the CRLF case statement before the guard runs) and the docs over-promised it; plus two fail-safe scope.py footguns. - install.sh: remove the dead self-heal block (.gitattributes is the real fix); INSTALL.md now tells the truth (renormalize an existing CRLF checkout manually) - engine/scope.py: keep bullets under nested ### sub-headings (was dropping in-scope assets under default-deny); --in-scope/--out-of-scope are now repeatable flags so a target placed after them is no longer swallowed - commands/scope.md: inline example uses repeated --in-scope flags - .gitignore: anchor scope.md -> /scope.md so it no longer shadows commands/scope.md Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * feat(skills): add triggers frontmatter to offensive-osint offensive-osint lacked the `triggers:` field that osint-methodology already has and that the sibling Claude-OSINT repo's frontmatter lint requires (>=5). Added the 127-entry recon trigger list so the skill is internally consistent and Claude-OSINT can re-export it without CI failure. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * feat(install): install manifest + safe --uninstall + skip-identical (Phase 3) Coexistence with the sister bundle Claude-OSINT: each install now records a manifest of its footprint so uninstall removes only its own files and keeps any skill the other bundle still owns (the two shared recon skills). - install.sh: write ~/.claude/.skill-manifests/claude-bughunter.txt; --uninstall removes tracked skills/commands/hunt.sh + rc line but keeps shared-owned skills; skip re-copying a skill/command already present and identical (no last-installer-wins clobber) - INSTALL.md: replace the unsafe rm -rf ~/.claude/skills advice with --uninstall - README: document safe coexistence with Claude-OSINT Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * feat(cbh): recon->hunt manifest contract (integration Phase 4) The data handoff between recon and hunt: cbh recon emits recon/<target>/manifest.json, which cbh surface and the hunt scaffold consume. Turns the conceptual 'OSINT recons, BugHunter hunts' into a real, tested contract. - cbh recon: emit manifest.json (schema 1.0) -- assets + ranked_surface (P1/P2/Kill via host heuristic + classify) + counts; secrets[]/identity_fabric{} left for the offensive-osint skill's deeper probes to append (shared-container design) - cbh surface <target>: new subcommand -- read the manifest, print the ranked surface - hunt.sh: ingest a manifest (2nd arg / $HUNT_MANIFEST / ./recon/<target>/) to seed scope.md (live hosts) + notes.md (ranked surface) - docs/recon-manifest.md + recon-manifest.schema.json: the contract spec - docs/cbh-cli.md, commands/surface.md: document the surface subcommand + manifest Round-trip verified end-to-end (synthetic + a real 'cbh recon example.com'). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * feat(dx): P2 backlog — --no-shell flag, supply-chain trust section, feedback template - install.sh: add --no-shell (don't touch shell rc; print the source line to add); the default path now also echoes the exact line it appended (T2) - SECURITY.md: 'Verifying what you install' supply-chain trust section — leans on the existing skill-lint CI, with an explicit 'what this does NOT prove' disclaimer and pin-to-reviewed-commit guidance, no malware-scanner theater (T13) - add .github/ISSUE_TEMPLATE/dx_feedback.yml — low-friction 'skill didn't trigger / first-hunt' DX feedback form (T14) Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Brian Richter <richter83.star@gmail.com> Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com> | 4 months ago | |
feat: add hunt-fintech-graphql skill + doc-count CI, refresh counts (#61) New `hunt-fintech-graphql` skill (money-movement GraphQL mutations, ledger/balance IDOR, decimal-precision abuse, idempotency-key replay, nested field-level KYC/PII authz, admin-override mass assignment; defers generic GraphQL to hunt-graphql). Adds scripts/check_doc_counts.py + doc-count-check.yml CI to guard skill/command counts against drift, and refreshes every doc count to disk truth (83 skills, 58 hunt-*, 15 commands). Contributed by @prithvee07. Original PR title described only the doc edits; rewritten to reflect the full scope. | 3 months ago | |
Skills: 71 → 82 — add 9, expand 11, dedup dispatch, portability fixes (#45) * feat(skills): 71 -> 82 skills — add 9, expand 11, dedup dispatch, YAML/portability fixes New skills (9): hunt-jwt-crypto, hunt-rag-vector, hunt-shadow-api, hunt-captcha-bypass, hunt-clickjacking, hunt-html-injection, hunt-forgot-password, hunt-exceptional-conditions, ios-redteam-pipeline. Plus hunt-spa-api, recon-scope-triage. Expanded 11 existing skills with verified technique content (sqli, ssti, xss, csrf, rce, session, mfa-bypass, race-condition, cache-poison, brute-force, api-misconfig). Dedup (description-scoping only, bodies untouched): jwt-crypto as JWT-crypto owner; differentiate bb-local-toolkit from bug-bounty; scope sqli/nosqli, auth-bypass/saml, cache-poison/host-header, report-writing+security-arsenal/triage-validation, spa-api. Fix 2 YAML parse blockers (clickjacking, html-injection); genericize lab/harness coupling and client-neutral placeholders for cross-harness portability; fold hunt-ai-attacks into hunt-llm-ai; regenerate cbh skill index (82 skills). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * docs: update skill count 71 -> 82 in banner and README Banner (banner-v2.svg): 3 count references (hero text, big SKILLS number, harness caption) — number-only change, layout/width preserved, valid XML. README: 5 references incl. vendored-foundation '8 of 71' -> '8 of 82'. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * docs: sync skill count 71 -> 82 across docs + regenerate catalog - Regenerate docs/skills.md catalog (82 skills, all 11 new rows, grouped) via gen_skill_catalog.py; add ios-redteam-pipeline to the enterprise group set. - Update count refs: INSTALL.md, docs/index.md (48->57 hunt sub-stack), docs/multi-harness.md (x3), docs/credits.md, docs/architecture.md. - Recompute origin/community/vendored breakdown 43/20/8 -> 54/20/8 (all 11 additions are original work); credits total 71 -> 82. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * docs(architecture): slot 11 new skills into phase tables - Add 9 new hunt-* to Phase 3 hunt list (48 -> 57) - Add ios-redteam-pipeline to enterprise layer (7 -> 8); update iOS 'gap' note - Add recon-scope-triage to Phase 2 Recon - Sub-counts synced: 57-skill hunt sub-stack, 8-skill enterprise layer Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Sachin Sharma <sachin.lucideus@gmail.com> Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com> | 3 months ago | |
fix(packaging): keep the standalone skill and report index current (#100) | 6 days ago | |
chore(star-history): refresh daily instead of weekly + guard against truncated fetch (#62) The self-hosted star chart already auto-refreshes, but only weekly (Monday cron), so the baked count lagged up to 7 days behind reality. Switch the cron to daily (04:17 UTC) so it's never more than ~1 day stale. Still fully self-hosted — no external service, same auto-commit-if-changed step. Also harden fetch_stars(): it ignored the `gh` exit code and treated any empty page as end-of-data, so a rate-limited/transient failure mid-pagination would silently truncate the series and commit a chart with a LOWER count. Daily runs multiply that exposure. Now a non-zero `gh api` exit aborts the run, leaving the committed SVGs untouched rather than regressing the count. README caption updated (weekly → daily) to match. Co-authored-by: Sachin Sharma <elementalsoul@Sachins-MacBook-Pro.local> Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com> | 3 months ago | |
fix(scaffold): seed the PowerShell engagement from recon natively (#107) | 6 days ago | |
fix(scaffold): keep Bash engagement creation usable with malformed recon manifests (#146) | 5 days ago | |
feat(install): add native Windows PowerShell installers (#49) Every .sh installer now has a PowerShell counterpart so the bundle installs natively on Windows without WSL: - scripts/hunt.ps1 (port of hunt.sh) - scripts/install.ps1 (port of install.sh) - scripts/install-community-skills.ps1 (port of install-community-skills.sh) Same behavior, same flags (hyphen-style: -All, -Agents, -Hermes, -BurpMcp, -NoProfile, -Uninstall), same manifest-driven uninstall. hunt.ps1 is dot-sourced from the PowerShell \$PROFILE (the Windows equivalent of the .zshrc/.bashrc source line). BOM-free UTF-8 writes. .gitattributes pins *.ps1 to LF alongside *.sh/*.py. Docs (README, INSTALL, USAGE, multi-harness, SECURITY, index, credits, CHANGELOG) updated to show both macOS/Linux and Windows paths. macOS/Linux unchanged. | 3 months ago | |
Generalize install path: ~/Research → ~/security-research The previous ~/Research was the maintainer's personal folder name. Replacing with the descriptive, generic ~/security-research across README, INSTALL, USAGE, the community-skills installer, and the two skill files that referenced it. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> | 6 months ago | |
feat(mcp): wire the HackerOne MCP into all harnesses like Burp (#165) The HackerOne MCP (#162) only installed into Claude Code; Burp wires into OpenCode/Codex/Hermes/AntiGravity via setup_harness_mcp.py. This brings parity. - setup_harness_mcp.py: parametrize the four writers by server name (default "burp", so existing behavior/tests are unchanged); add hackerone_source() (python3 + in-repo server path, carries HACKERONE_API_* env only if set); add --burp/--hackerone selection (default burp). A burp-discovery miss now warns and continues when hackerone is also requested, instead of aborting. - install.sh / install.ps1: add --hackerone-mcp / -HackeroneMcp, passing the server-selection flags through; both servers can be wired in one run. - docs/multi-harness.md: HackerOne MCP section mirroring the Burp one. - tests: test_mcp_hackerone_harness.py (source resolution + keyed wiring). Co-authored-by: Sachin Sharma <sachin.lucideus@gmail.com> Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com> | 4 days ago | |
feat(mcp): wire the HackerOne MCP into all harnesses like Burp (#165) The HackerOne MCP (#162) only installed into Claude Code; Burp wires into OpenCode/Codex/Hermes/AntiGravity via setup_harness_mcp.py. This brings parity. - setup_harness_mcp.py: parametrize the four writers by server name (default "burp", so existing behavior/tests are unchanged); add hackerone_source() (python3 + in-repo server path, carries HACKERONE_API_* env only if set); add --burp/--hackerone selection (default burp). A burp-discovery miss now warns and continues when hackerone is also requested, instead of aborting. - install.sh / install.ps1: add --hackerone-mcp / -HackeroneMcp, passing the server-selection flags through; both servers can be wired in one run. - docs/multi-harness.md: HackerOne MCP section mirroring the Burp one. - tests: test_mcp_hackerone_harness.py (source resolution + keyed wiring). Co-authored-by: Sachin Sharma <sachin.lucideus@gmail.com> Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com> | 4 days ago | |
fix(skills): replace literal NUL in hunt-auth-bypass; lint rejects control chars (#133) The payload example in skills/hunt-auth-bypass/SKILL.md carried a raw 0x00 byte (offset 23277, line 359) where the visible text \x00 was intended. That single byte makes the file binary to ripgrep and file(1), so repo-wide searches skip the skill silently, while the loader, GitHub rendering, git diff and CI all stay green - which is why it survived since c42e50d7 (2026-05-25). - write the escape as visible text, so rg finds the file again and file(1) reports UTF-8 text - add a control-character check to scripts/lint_skills.py so the class cannot come back; it reports line and byte offset of the offender Verified: scripts/lint_skills.py -> 0 errors, 14 warnings (unchanged); the new rule flags the pre-fix file as hunt-auth-bypass:359. | 6 days ago | |
Phase 3.2: CISA KEV coverage refresh script (+ workflow template) Closes the "currency drift" gap. Pulls the CISA Known Exploited Vulnerabilities catalog, greps the repo for indexed CVE IDs, surfaces missing entries from vendors that map to our skill stack. Components shipped: - scripts/refresh-cve-index.py — pulls KEV JSON, scope-filters by vendor (Apache, Atlassian, Cisco, Citrix, F5, Fortinet, GitLab, Ivanti, Jenkins, Microsoft, Oracle, Palo Alto, Pulse, Red Hat, SonicWall, Spring, VMware, WordPress, Drupal, Joomla, PHP, Adobe). Out-of-scope filter for browser/kernel/mobile-internal CVEs that don't map to bug-bounty / external-red-team work. --markdown flag for the snapshot file format. - docs/cve-coverage.md — committed snapshot (KEV 2026.05.15: 1592 total vulns, 91 indexed in this repo, 54 in-scope gaps in 180-day lookback). - docs/automation/cve-refresh.yml.template — GitHub Actions workflow template. To install, move it to .github/workflows/ after granting the workflow scope: gh auth refresh -h github.com -s workflow mkdir -p .github/workflows mv docs/automation/cve-refresh.yml.template .github/workflows/cve-refresh.yml git add .github/workflows/cve-refresh.yml && git commit && git push The workflow runs weekly (Mondays 14:00 UTC), opens/updates a single tracking issue if gaps are detected, labels it cve-coverage + automation. First run found 91 already-indexed CVEs covering Log4Shell, ToolShell, Jenkins CVE-2024-23897, Spring CVE-2022-22963, Apache CVE-2021-41773 plus the ~85 cited across the Pattern Library docs from Phase 3. Highest-priority coverage gaps detected (from KEV last 180 days): CVE-2026-42897 (Microsoft Exchange XSS), CVE-2026-20182 (Cisco SD-WAN auth bypass), CVE-2026-6973 (Ivanti EPMM), CVE-2026-0300 (PAN-OS), CVE-2026-34197 (Apache ActiveMQ), CVE-2025-53521 (F5 BIG-IP). Each is a candidate for a new disclosed-reports / skill entry. Currency drift problem solved: the repo can now answer "are we current with CISA KEV?" every Monday, not "we were current when we last looked". Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> | 6 months ago | |
fix(hooks): scan staged content before allowing a commit (#112) | 6 days ago | |
feat(mcp): wire the HackerOne MCP into all harnesses like Burp (#165) The HackerOne MCP (#162) only installed into Claude Code; Burp wires into OpenCode/Codex/Hermes/AntiGravity via setup_harness_mcp.py. This brings parity. - setup_harness_mcp.py: parametrize the four writers by server name (default "burp", so existing behavior/tests are unchanged); add hackerone_source() (python3 + in-repo server path, carries HACKERONE_API_* env only if set); add --burp/--hackerone selection (default burp). A burp-discovery miss now warns and continues when hackerone is also requested, instead of aborting. - install.sh / install.ps1: add --hackerone-mcp / -HackeroneMcp, passing the server-selection flags through; both servers can be wired in one run. - docs/multi-harness.md: HackerOne MCP section mirroring the Burp one. - tests: test_mcp_hackerone_harness.py (source resolution + keyed wiring). Co-authored-by: Sachin Sharma <sachin.lucideus@gmail.com> Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com> | 4 days ago | |
fix(hooks): scan staged content before allowing a commit (#112) | 6 days ago | |
fix(engine,cli): cross-platform Windows support, utf-8 file writing, and triage test suite (#160) | 4 days ago |
| Files | Last commit | Last update |
|---|---|---|
| 2 months ago | ||
| 4 months ago | ||
| 3 months ago | ||
| 3 months ago | ||
| 6 days ago | ||
| 3 months ago | ||
| 6 days ago | ||
| 5 days ago | ||
| 3 months ago | ||
| 6 months ago | ||
| 4 days ago | ||
| 4 days ago | ||
| 6 days ago | ||
| 6 months ago | ||
| 6 days ago | ||
| 4 days ago | ||
| 6 days ago | ||
| 4 days ago |