pentesting 0.73.1 → 0.73.3
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +3 -14
- package/dist/agent-tool-JEFUBDZE.js +989 -0
- package/dist/{chunk-EUWDAGHZ.js → chunk-BKWCGMSV.js} +879 -427
- package/dist/{chunk-YFDJI3GO.js → chunk-GLO6TOJN.js} +2 -0
- package/dist/{chunk-ZQAVMACI.js → chunk-UB7RW6LM.js} +267 -153
- package/dist/main.js +1378 -198
- package/dist/{persistence-SNUMO4WG.js → persistence-2WKQHGOL.js} +2 -2
- package/dist/{process-registry-GSHEX2LT.js → process-registry-QIW7ZIUT.js} +1 -1
- package/dist/prompts/main-agent.md +35 -1
- package/dist/prompts/strategist-system.md +34 -0
- package/package.json +1 -1
- package/dist/agent-tool-COMG67ES.js +0 -256
package/README.md
CHANGED
|
@@ -1,12 +1,12 @@
|
|
|
1
1
|
<div align="center">
|
|
2
2
|
|
|
3
|
-
<img src="https://api.iconify.design/game-icons:fizzing-flask.svg?color=%
|
|
3
|
+
<img src="https://api.iconify.design/game-icons:fizzing-flask.svg?color=%232496ED" width="80" height="80" alt="Pentesting Agent" />
|
|
4
4
|
|
|
5
5
|
# pentesting
|
|
6
6
|
> **Autonomous Offensive Security AI Agent**
|
|
7
7
|
|
|
8
|
-
[](https://www.npmjs.org/package/pentesting)
|
|
9
|
+
[](https://hub.docker.com/r/agnusdei1207/pentesting)
|
|
10
10
|
|
|
11
11
|
</div>
|
|
12
12
|
|
|
@@ -32,17 +32,6 @@
|
|
|
32
32
|
|
|
33
33
|
Pentesting support tool. Can autonomously execute network penetration tests or assist with generic Capture The Flag (CTF) challenges (such as Reverse Engineering, Cryptography, and binary analysis) without requiring a specific network target.
|
|
34
34
|
|
|
35
|
-
## Architecture Notes
|
|
36
|
-
|
|
37
|
-
- User input is preprocessed by a dedicated input processor LLM before the main loop acts on it.
|
|
38
|
-
- Durable engagement guidance, sensitive data handling rules, and reusable operator constraints are merged into `.pentesting/memory/policy.md`.
|
|
39
|
-
- Both the strategist and the main prompt builder read that policy document every turn.
|
|
40
|
-
- Each completed turn is compressed into `.pentesting/turns/{N}-memory.md`, with provenance metadata describing who wrote it and what sources were used.
|
|
41
|
-
- Automatically maintained LLM documents are intentionally small in number: bounded turn memories, one `policy.md`, one merged `persistent-knowledge.json`, and on-demand reports only.
|
|
42
|
-
- Interactive prompts are brokered through a single active input slot in the TUI. Additional prompts wait in a hidden queue and are promoted one at a time.
|
|
43
|
-
|
|
44
|
-
---
|
|
45
|
-
|
|
46
35
|
## Quick Start
|
|
47
36
|
|
|
48
37
|
### z.ai — GLM Coding Plan Max (Recommended)
|