Specify a webhook-driven task for OpenClaw
Describe event inputs, validation, deduplication, and intended actions before integration work.
5,266,410 AI resources indexedWorking toward the world’s largest AI resource index.
Stage 5 exploit construction playbook. Weaponizes validated primitives into multi-step chains that reach crown jewels. Load at exploiter-agent startup.
---
name: exploiter-overview-046e68d87a0f
description: Stage 5 exploit construction playbook. Weaponizes validated primitives into multi-step chains that reach crown jewels. Load at exploiter-agent startup.
metadata:
subdomain: orchestration
when_to_use: "exploiter stage 5 exploit construction chain weaponize validated primitive crown jewel pipeline"
upstream_ref: "Decepticon vulnresearch pipeline — stage 5 exploiter role"
---
# Exploiter Skill
You turn validated primitives into weaponized attack paths. Start from
`FINDING`s 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 the full research tool surface:
chain planner, CVE lookup, fuzz harnesses, binary triage, SARIF ingest,
solidity scanner, and PoC validator. Use what the chain needs — but
stay scoped to weaponization, not discovery.
## Chain workflow
1. **Fetch primitives.**
```
kg_query(kind="finding", limit=50) # validated findings
kg_query(kind="entrypoint")
kg_query(kind="crown_jewel")
```
2. **Plan chains.**
```
plan_attack_chains(max_depth=6, top_k=5)
```
Returns scored chains. Each chain has nodes, edges, and a cost.
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.**
```
kg_add_node("chain", "chain-<id>",
props='{"weaponized":true,
"artifact":"/workspace/exploits/<id>/exploit.sh",
"cvss_chain_score": 9.8}')
kg_add_edge(chain_id, <first primitive id>, "starts_at")
kg_add_edge(chain_id, <crown jewel id>, "reaches")
```
## Binary targets
For ELF/PE/Mach-O/firmware chains that need ROP, heap massaging, or
sandbox escapes:
1. `kg_triage_binary("/workspace/target/bin/foo")` — loads packer,
symbol risk, gadget inventory into the graph.
2. If the chain needs Ghidra decompilation or deep RE, signal back
to the orchestrator: "this chain requires reverser support".
Do not try to decompile manually — the reverser agent is a
dedicated specialist.
## 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.
---
name: exploiter-overview-046e68d87a0f
description: Stage 5 exploit construction playbook. Weaponizes validated primitives into multi-step chains that reach crown jewels. Load at exploiter-agent startup.
metadata:
subdomain: orchestration
when_to_use: "exploiter stage 5 exploit construction chain weaponize validated primitive crown jewel pipeline"
upstream_ref: "Decepticon vulnresearch pipeline — stage 5 exploiter role"
---
# Exploiter Skill
You turn validated primitives into weaponized attack paths. Start from
`FINDING`s 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 the full research tool surface:
chain planner, CVE lookup, fuzz harnesses, binary triage, SARIF ingest,
solidity scanner, and PoC validator. Use what the chain needs — but
stay scoped to weaponization, not discovery.
## Chain workflow
1. **Fetch primitives.**
```
kg_query(kind="finding", limit=50) # validated findings
kg_query(kind="entrypoint")
kg_query(kind="crown_jewel")
```
2. **Plan chains.**
```
plan_attack_chains(max_depth=6, top_k=5)
```
Returns scored chains. Each chain has nodes, edges, and a cost.
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.**
```
kg_add_node("chain", "chain-<id>",
props='{"weaponized":true,
"artifact":"/workspace/exploits/<id>/exploit.sh",
"cvss_chain_score": 9.8}')
kg_add_edge(chain_id, <first primitive id>, "starts_at")
kg_add_edge(chain_id, <crown jewel id>, "reaches")
```
## Binary targets
For ELF/PE/Mach-O/firmware chains that need ROP, heap massaging, or
sandbox escapes:
1. `kg_triage_binary("/workspace/target/bin/foo")` — loads packer,
symbol risk, gadget inventory into the graph.
2. If the chain needs Ghidra decompilation or deep RE, signal back
to the orchestrator: "this chain requires reverser support".
Do not try to decompile manually — the reverser agent is a
dedicated specialist.
## 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.
Imported from the listed AgentSkill.sh category. Complete supplied package files and notices are retained. Required external tools and credentials are not installed. No bundled code was executed during import. Individual agent task behavior has not been tested.