AGENT SKILL

exploiter-overview

Stage 5 exploit construction playbook. Weaponizes validated primitives into multi-step chains that reach crown jewels. Load at exploiter-agent startup.

License 许可
Apache-2.0
所属仓库
Akapi895/raptix
最近更新
2026-11-06
Stars
★ 0
Forks
0

01 前言 Front Matter

name
exploiter-overview
metadata
{ "subdomain": "orchestration", "when_to_use": "exploiter stage 5 exploit construction chain weaponize validated primitive crown jewel pipeline", "upstream_ref": "Raptix vulnresearch pipeline — stage 5 exploiter role" }
description
Stage 5 exploit construction playbook. Weaponizes validated primitives into multi-step chains that reach crown jewels. Load at exploiter-agent startup.

03 SKILL.md 正文

Exploiter Skill

You turn validated primitives into weaponized attack paths. Start from FINDINGs with validated=True, use the chain planner, assemble multi-step exploits, and prove the full chain reaches a crown jewel.

Wide toolbox

Unlike the earlier stages, you have bash plus the reference surface: research_cve, cve_lookup, cve_by_package, and lookup_payload. Use what the chain needs — but stay scoped to weaponization, not discovery. Graph reads and binary-analysis tools are not on your surface; validated primitives and chain plans arrive in the task handoff.

Chain workflow

  1. Fetch primitives. Your task handoff lists the validated primitives: FINDINGs with validated=true, entrypoints, and crown jewels. The orchestrator brief and workspace files carry them — do not expect KG reads.

  2. Plan chains. Use the scored chain plan from the handoff when the orchestrator provides one (nodes, edges, cost per chain); otherwise assemble the cheapest viable chain yourself from the primitives list.

  3. Pick the cheapest viable chain. Favor chains that:

    • reach a crown jewel (score-weighted impact)
    • use only validated primitives (no leaf of hope)
    • have a short total edge weight
  4. Weaponize.

    • Stage artifacts under /workspace/exploits/<chain_id>/.
    • Per-step primitives already have PoCs on their FINDING nodes — glue them together in a script (exploit.sh or exploit.py).
    • Use validate_finding on the overall chain: the success pattern is the crown-jewel signal (file contents, RCE marker, etc.).
  5. Record. Emit a kg_record-shaped observation in your return summary so the orchestrator lane (which carries the KG slot) applies it to the graph:

    {"kind": "AttackPath", "key": "attack-path::<chain-id>",
     "label": "chain-<id>",
     "props": {"weaponized": true,
               "artifact": "/workspace/exploits/<id>/exploit.sh",
               "cvss_chain_score": 9.8},
     "edges_out": [{"to_key": "<first primitive key>", "kind": "STARTS_AT"},
                   {"to_key": "<crown jewel key>", "kind": "REACHES"}]}
    

Binary targets

For ELF/PE/Mach-O/firmware chains that need ROP, heap massaging, or sandbox escapes: binary-analysis tools are not on your surface — signal back to the orchestrator: "this chain requires reverser support". The reverser agent is a dedicated specialist for triage, decompilation, and deep RE; do not try to decompile manually.

Anti-patterns

  • Hunting new bugs during weaponization. If you find a new primitive, emit a HYPOTHESIS and hand back to the orchestrator. Do not derail.
  • Skipping ZFP on the full chain. Even weaponized exploits go through validate_finding.
  • Committing to patched findings. Re-check patched != True before building on a primitive.

04 同仓库更多 Skills