[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7EGT4SB8CG3HTJ31B5JWW2F":3,"skill-same:01M7EGT4SB8CG3HTJ31B5JWW2F":46},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":24,"commitSha":25,"skillDirectory":26,"forksCount":27,"repositoryHtmlUrl":28,"repositoryHomepageUrl":6,"ownerHtmlUrl":29,"ownerBlogUrl":6,"ownerDescription":30,"updatedAt":31,"install":32,"resources":45},"01M7EGT4SB8CG3HTJ31B5JWW2F","ctf-workflow",null,"Autonomous CTF \u002F boot-to-root campaign driver. Runs a box end to end with no operator approvals - the deterministic driver (scripts\u002Fcampaign.py) owns pass state, generates the killchain board from recon, and prints the exact next action (Skill + tool) every turn. Use when handed a box\u002FIP to own end to end, \"run the ctf workflow\", \"root this box\", \"foothold to root\". Single agent, wiki-first, tool-first. OSINT is OFF unless you invoke with an osint argument. Delegates box recipes to ctf-box.","lza6\u002FArmor-piercing-system-prompts","lza6","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F128880206?v=4","01M7EB3K7WPX5T28EJ2NVHE6EG","Chinese-tingfeng","01M7EB3K97E1KQ16QWKJMFNGX8",11,"MIT","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F04-安全渗透\u002Fctf-workflow\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: ctf-workflow\ndescription: Autonomous CTF \u002F boot-to-root campaign driver. Runs a box end to end with no operator approvals - the deterministic driver (scripts\u002Fcampaign.py) owns pass state, generates the killchain board from recon, and prints the exact next action (Skill + tool) every turn. Use when handed a box\u002FIP to own end to end, \"run the ctf workflow\", \"root this box\", \"foothold to root\". Single agent, wiki-first, tool-first. OSINT is OFF unless you invoke with an osint argument. Delegates box recipes to ctf-box.\n---\n\n# ctf-workflow\n\nThe driver is the plan. Run one command, do exactly what it prints, record the result, repeat. The\ngates are enforced by `scripts\u002Fcampaign.py`.\n\n## The loop\n\n```\npython3 scripts\u002Fcampaign.py next\npython3 scripts\u002Fcampaign.py note \u003Crow> --arsenal \u003Cslug>\npython3 scripts\u002Fcampaign.py done \u003Crow> --poc \u003Cimg> --kind req   # | --dead R | --park Q | --find F\n```\n\nEvery `next` now prints an `APPROACH:`\u002F`AVOID:`\u002F`REFS:` block for the served row (the distilled\nctf-box method for that vuln class) - read it before acting.\n\n## Start \u002F resume\n\n1. `python3 scripts\u002Fcampaign.py init --type ctf` - validates `scope.md` (for a box, scope is just\n   the target IP\u002Fhost) + envelope.\n2. **OSINT is OFF by default** for CTF (pass 0 is skipped). Only run it if this skill was invoked\n   with an explicit `osint` argument - a box's answer is on the box, not in Wayback.\n3. Passes 1-3 feed `state.md` - rustscan\u002Fnmap + web enum. Read every service banner, page source and\n   config end to end.\n4. `python3 scripts\u002Fcampaign.py board` - writes the 4a foothold rows; once an asset is a foothold,\n   re-run `board` and it seeds the 4b privesc rows (pspy\u002Flinpeas auto + the manual checklist) for\n   that asset.\n5. Enter the loop, depth-first.\n6. **When a foothold lands** (a reverse shell in a tmux window via `vm-scan.sh --win shell`, or a\n   meterpreter\u002Fmsfconsole session via `--win msf`), record it: `python3 scripts\u002Fcampaign.py foothold\n   \u003Ctarget> --win shell` (or `--win msf`; or ride it on the closing find with `done ... --win`). The\n   driver flips the asset's `state.md` row to `access=foothold` and routes its 4b privesc rows through\n   `vm-rsh --win \u003Cwin>` (persistent session + operator visibility past foothold), and prints `tmux\n   attach -t \u003Ceng>` for manual takeover. msf itself is operator-attach \u002F drop-to-shell, not\n   `vm-rsh`-driven (its wrapper frames a bash shell, not the `msf6 >` REPL). **After recording a\n   foothold, re-run `python3 scripts\u002Fcampaign.py board` so the 4b privesc rows are seeded** - `next`\n   will not surface them until you do.\n7. **Web RCE -> a real shell, THEN stabilize -- do not ride one-liners (recurring drift).** The\n   moment code-exec lands (a web-RCE primitive, an LFI->session-poison, a deser gadget), STOP\n   hand-poking one-shot payloads: (a) **catch it with a real handler by default** - `msfconsole`'s\n   `exploit\u002Fmulti\u002Fhandler` (meterpreter first; a plain `shell_reverse_tcp` is the backup when\n   meterpreter is blocked, e.g. Windows\u002FEDR) (record via `campaign.py foothold \u003Ctarget> --win msf`,\n   step 6). Reserve a raw `nc -lvnp` listener for when msf is unavailable: a raw nc pane's `Ctrl-C`\n   kills the LISTENER (dropping the shell back to your own prompt, the false-root attacker-prompt\n   trap) and it has no session management, while\n   meterpreter also carries `post\u002Fmulti\u002Frecon`\u002F`local_exploit_suggester` escalation modules and\n   built-in file transfer. (b) **Before picking the LPORT, test target egress on common ports\n   (80\u002F443\u002F53)** - high ports like 4444 are frequently filtered, so pick an egress-allowed LPORT. (`scripts\u002Fvm-handler.sh \u003Ceng> \u003Clhost>` picks a free egress port and launches the handler for you, printing the LPORT.)\n   (c) If you did fall back to raw nc, **stabilize it immediately** with `bash\n   scripts\u002Fvm-stabilize.sh --win shell \u003Ceng>` (pty + job control + window size). (d) Then record the\n   foothold (step 6) and drive with `vm-rsh`. An unstabilized nc shell (no job control, mid-line\n   wrapping) is what makes post-ex drift back into one-liners. Full discipline: `Skill(ctf-box)`\n   Phase 3 (Deliver). The `recon-capture` hook fires this reflex once on the first service-account\n   `id`.\n\n## Box recipes\n\n`ctf-workflow` owns pass sequencing and the board; the per-service exploitation recipes live in\n`Skill(ctf-box)`, which the driver hands off to - do not duplicate them here. Route a fingerprinted\nservice to its `Skill(hunt-*)` as the board names it; use `Skill(ctf-category)` for a standalone\nchallenge (pwn\u002Frev\u002Fcrypto\u002Fforensics\u002Fstego).\n\n## Browser observation\n\nA box is VPN-boxed, so the local `chrome-devtools` MCP browser cannot reach it - use the **VM-side\nbrowser** (`scripts\u002Fbrowser.sh \u003Curl>` \u002F `capture.sh web`) to render a JS-heavy web service and read\nits DOM + network requests (the rendered XHR\u002Ffetch calls reveal API routes a `curl` crawl misses -\noften the intended path). Rendered screenshots of the flag\u002Fexploited state are valid `web` PoCs.\n\nA service needing a **manual login \u002F MFA \u002F CAPTCHA** the agent can't do headlessly -> `Skill(chrome-devtools-browser)`: a VISIBLE chromium on the VM desktop (`scripts\u002Fbrowser-visible.sh`) the operator drives, observed live via the chrome-devtools MCP.\n\n## Gates\n\nG1 arsenal-first, G2 skill-first, G3 typed evidence (a flag on screen is a valid `web` PoC), G8\ntool-first. Privesc always includes pspy + linpeas\u002Fwinpeas - the board seeds these as 4b rows once\nan asset is recorded as a foothold (re-run `board` after `foothold`\u002F`done --win` to seed them).\n\nOnce a foothold and a working escalation vector are identified, hand the exploit compile +\nescalation run to a sub-agent via `Skill(delegate)` - checklist, model choice, and the\nmandatory false-root\u002Ffalse-RCE `hostname`+uid guardrail all live there; keep the main agent\ndriving the board.\n\nPrefer a clean post-ex command channel over driving `msf sessions -c` (delayed output, quoting\npain): a webshell writing enum output to a web-served file then `curl` it, or a single-tool\nread\u002Fdecrypt done locally on already-exfiltrated data.\n\n## Autonomy\n\nNo approvals. The verifier is optional for CTF (a captured flag self-verifies). Both flags captured ->\nset `## STATUS: SOLVED` in state.md, then the driver prints the close-out chain.\n\n## Discipline\n\n- Do NOT invoke `superpowers:brainstorming`\u002F`writing-plans` mid-box; keep no parallel task list.\n- One agent. Read service output whole - the foothold hides in the handler a grep skips.\n- Reuse captured creds across hosts before researching new ones.\n- Long\u002Fobserved tools (sqlmap, big scans) go in a NAMED tmux window (`scripts\u002Fvm-scan.sh`), never a\n  blind background pipe -- you must WATCH a scanner that can trip a target's rate-limiter\u002Fban.\n- Capture the flag\u002Froot state to `poc\u002F` AS IT LANDS (`scripts\u002Fcapture.sh`), even on a curl\u002Fssh-only\n  box -- a transient exploited state cannot be re-shot after the turn.\n\n## Close-out\n\nRun the printed chain: `Skill(walkthrough)` -> `Skill(learn)`.\n\n## If the driver is unavailable\n\nManual fallback: read `Approach.md`, take the top open row, run its wiki lookup then its hunt skill\nor `Skill(ctf-box)`, capture evidence, mark `[x]`; on exhaustion one `Deadends.md` line + `[!]`.\n","# ctf-workflow\n\nThe driver is the plan. Run one command, do exactly what it prints, record the result, repeat. The\ngates are enforced by `scripts\u002Fcampaign.py`.\n\n## The loop\n\n```\npython3 scripts\u002Fcampaign.py next\npython3 scripts\u002Fcampaign.py note \u003Crow> --arsenal \u003Cslug>\npython3 scripts\u002Fcampaign.py done \u003Crow> --poc \u003Cimg> --kind req   # | --dead R | --park Q | --find F\n```\n\nEvery `next` now prints an `APPROACH:`\u002F`AVOID:`\u002F`REFS:` block for the served row (the distilled\nctf-box method for that vuln class) - read it before acting.\n\n## Start \u002F resume\n\n1. `python3 scripts\u002Fcampaign.py init --type ctf` - validates `scope.md` (for a box, scope is just\n   the target IP\u002Fhost) + envelope.\n2. **OSINT is OFF by default** for CTF (pass 0 is skipped). Only run it if this skill was invoked\n   with an explicit `osint` argument - a box's answer is on the box, not in Wayback.\n3. Passes 1-3 feed `state.md` - rustscan\u002Fnmap + web enum. Read every service banner, page source and\n   config end to end.\n4. `python3 scripts\u002Fcampaign.py board` - writes the 4a foothold rows; once an asset is a foothold,\n   re-run `board` and it seeds the 4b privesc rows (pspy\u002Flinpeas auto + the manual checklist) for\n   that asset.\n5. Enter the loop, depth-first.\n6. **When a foothold lands** (a reverse shell in a tmux window via `vm-scan.sh --win shell`, or a\n   meterpreter\u002Fmsfconsole session via `--win msf`), record it: `python3 scripts\u002Fcampaign.py foothold\n   \u003Ctarget> --win shell` (or `--win msf`; or ride it on the closing find with `done ... --win`). The\n   driver flips the asset's `state.md` row to `access=foothold` and routes its 4b privesc rows through\n   `vm-rsh --win \u003Cwin>` (persistent session + operator visibility past foothold), and prints `tmux\n   attach -t \u003Ceng>` for manual takeover. msf itself is operator-attach \u002F drop-to-shell, not\n   `vm-rsh`-driven (its wrapper frames a bash shell, not the `msf6 >` REPL). **After recording a\n   foothold, re-run `python3 scripts\u002Fcampaign.py board` so the 4b privesc rows are seeded** - `next`\n   will not surface them until you do.\n7. **Web RCE -> a real shell, THEN stabilize -- do not ride one-liners (recurring drift).** The\n   moment code-exec lands (a web-RCE primitive, an LFI->session-poison, a deser gadget), STOP\n   hand-poking one-shot payloads: (a) **catch it with a real handler by default** - `msfconsole`'s\n   `exploit\u002Fmulti\u002Fhandler` (meterpreter first; a plain `shell_reverse_tcp` is the backup when\n   meterpreter is blocked, e.g. Windows\u002FEDR) (record via `campaign.py foothold \u003Ctarget> --win msf`,\n   step 6). Reserve a raw `nc -lvnp` listener for when msf is unavailable: a raw nc pane's `Ctrl-C`\n   kills the LISTENER (dropping the shell back to your own prompt, the false-root attacker-prompt\n   trap) and it has no session management, while\n   meterpreter also carries `post\u002Fmulti\u002Frecon`\u002F`local_exploit_suggester` escalation modules and\n   built-in file transfer. (b) **Before picking the LPORT, test target egress on common ports\n   (80\u002F443\u002F53)** - high ports like 4444 are frequently filtered, so pick an egress-allowed LPORT. (`scripts\u002Fvm-handler.sh \u003Ceng> \u003Clhost>` picks a free egress port and launches the handler for you, printing the LPORT.)\n   (c) If you did fall back to raw nc, **stabilize it immediately** with `bash\n   scripts\u002Fvm-stabilize.sh --win shell \u003Ceng>` (pty + job control + window size). (d) Then record the\n   foothold (step 6) and drive with `vm-rsh`. An unstabilized nc shell (no job control, mid-line\n   wrapping) is what makes post-ex drift back into one-liners. Full discipline: `Skill(ctf-box)`\n   Phase 3 (Deliver). The `recon-capture` hook fires this reflex once on the first service-account\n   `id`.\n\n## Box recipes\n\n`ctf-workflow` owns pass sequencing and the board; the per-service exploitation recipes live in\n`Skill(ctf-box)`, which the driver hands off to - do not duplicate them here. Route a fingerprinted\nservice to its `Skill(hunt-*)` as the board names it; use `Skill(ctf-category)` for a standalone\nchallenge (pwn\u002Frev\u002Fcrypto\u002Fforensics\u002Fstego).\n\n## Browser observation\n\nA box is VPN-boxed, so the local `chrome-devtools` MCP browser cannot reach it - use the **VM-side\nbrowser** (`scripts\u002Fbrowser.sh \u003Curl>` \u002F `capture.sh web`) to render a JS-heavy web service and read\nits DOM + network requests (the rendered XHR\u002Ffetch calls reveal API routes a `curl` crawl misses -\noften the intended path). Rendered screenshots of the flag\u002Fexploited state are valid `web` PoCs.\n\nA service needing a **manual login \u002F MFA \u002F CAPTCHA** the agent can't do headlessly -> `Skill(chrome-devtools-browser)`: a VISIBLE chromium on the VM desktop (`scripts\u002Fbrowser-visible.sh`) the operator drives, observed live via the chrome-devtools MCP.\n\n## Gates\n\nG1 arsenal-first, G2 skill-first, G3 typed evidence (a flag on screen is a valid `web` PoC), G8\ntool-first. Privesc always includes pspy + linpeas\u002Fwinpeas - the board seeds these as 4b rows once\nan asset is recorded as a foothold (re-run `board` after `foothold`\u002F`done --win` to seed them).\n\nOnce a foothold and a working escalation vector are identified, hand the exploit compile +\nescalation run to a sub-agent via `Skill(delegate)` - checklist, model choice, and the\nmandatory false-root\u002Ffalse-RCE `hostname`+uid guardrail all live there; keep the main agent\ndriving the board.\n\nPrefer a clean post-ex command channel over driving `msf sessions -c` (delayed output, quoting\npain): a webshell writing enum output to a web-served file then `curl` it, or a single-tool\nread\u002Fdecrypt done locally on already-exfiltrated data.\n\n## Autonomy\n\nNo approvals. The verifier is optional for CTF (a captured flag self-verifies). Both flags captured ->\nset `## STATUS: SOLVED` in state.md, then the driver prints the close-out chain.\n\n## Discipline\n\n- Do NOT invoke `superpowers:brainstorming`\u002F`writing-plans` mid-box; keep no parallel task list.\n- One agent. Read service output whole - the foothold hides in the handler a grep skips.\n- Reuse captured creds across hosts before researching new ones.\n- Long\u002Fobserved tools (sqlmap, big scans) go in a NAMED tmux window (`scripts\u002Fvm-scan.sh`), never a\n  blind background pipe -- you must WATCH a scanner that can trip a target's rate-limiter\u002Fban.\n- Capture the flag\u002Froot state to `poc\u002F` AS IT LANDS (`scripts\u002Fcapture.sh`), even on a curl\u002Fssh-only\n  box -- a transient exploited state cannot be re-shot after the turn.\n\n## Close-out\n\nRun the printed chain: `Skill(walkthrough)` -> `Skill(learn)`.\n\n## If the driver is unavailable\n\nManual fallback: read `Approach.md`, take the top open row, run its wiki lookup then its hunt skill\nor `Skill(ctf-box)`, capture evidence, mark `[x]`; on exhaustion one `Deadends.md` line + `[!]`.\n",{"name":5,"description":7},"main","d69096e239f2a89aabc2cd078c47664a9d3b1e11","04-安全渗透\u002Fctf-workflow",3,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts","https:\u002F\u002Fgithub.com\u002Flza6","个人vx：Tf00798\r\n个人✈：@TFGY999","2026-10-08 14:29:22",{"source":8,"skillName":5,"commands":33,"prompt":43,"zipAvailable":44},[34,37,40],{"tool":35,"command":36},"npx","npx skills add lza6\u002FArmor-piercing-system-prompts --skill ctf-workflow",{"tool":38,"command":39},"bunx","bunx skills add lza6\u002FArmor-piercing-system-prompts --skill ctf-workflow",{"tool":41,"command":42},"pnpm","pnpm dlx skills add lza6\u002FArmor-piercing-system-prompts --skill ctf-workflow","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F04-安全渗透\u002Fctf-workflow\u002FSKILL.md",true,[],[47,55,60,67,74,81],{"skillNo":48,"name":49,"nameZh":6,"description":50,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":51,"categories":52,"keywords":53,"publishedAt":54},"01M7FFG485KY83TD6HSC4YPCMC","obsidian-clipper-template-creator","Guide for creating templates for the Obsidian Web Clipper. Use when you want to create a new clipping template, understand available variables, or format clipped content.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobsidian-clipper-template-creator\u002FSKILL.md",[],[],"2026-11-16 10:05:34",{"skillNo":56,"name":49,"nameZh":6,"description":50,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":57,"categories":58,"keywords":59,"publishedAt":54},"01M7FFG0T6AZ0STEJPSB3CX84K","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobsidian-clipper-template-creator-antigravity-awesome-skills-main\u002FSKILL.md",[],[],{"skillNo":61,"name":62,"nameZh":6,"description":63,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":64,"categories":65,"keywords":66,"publishedAt":54},"01M7FFFY0QEM14SKWPDYGMXM2P","python-api","Programmatic Object-Detection-Metrics API guidance for BoundingBox, BoundingBoxes, Evaluator, enums, coordinate conversion, AP\u002FmAP semantics, and source-style import diagnostics.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobject-detection-metrics\u002Fsub-skills\u002Fpython-api\u002FSKILL.md",[],[],{"skillNo":68,"name":69,"nameZh":6,"description":70,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":71,"categories":72,"keywords":73,"publishedAt":54},"01M7FFFVZQZ2GJ0C0TGP7JHD20","file-evaluation","Folder-based PASCAL VOC AP\u002FmAP evaluation from ground-truth and detection text files.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobject-detection-metrics\u002Fsub-skills\u002Ffile-evaluation\u002FSKILL.md",[],[],{"skillNo":75,"name":76,"nameZh":6,"description":77,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":78,"categories":79,"keywords":80,"publishedAt":54},"01M7FFFSFNYSE77SJ6PE1CK2DA","object-detection-metrics","Guides legacy Object-Detection-Metrics PASCAL VOC AP\u002FmAP evaluation from detection text folders and direct Python API classes.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobject-detection-metrics\u002FSKILL.md",[],[],{"skillNo":82,"name":83,"nameZh":6,"description":84,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":85,"categories":86,"keywords":87,"publishedAt":54},"01M7FFFP0GCK1RVXBCPQ8APBMT","nx-workspace-patterns","Configure and optimize Nx monorepo workspaces. Use when setting up Nx, configuring project boundaries, optimizing build caching, or implementing affected commands.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fnx-workspace-patterns\u002FSKILL.md",[],[]]