# amruth-sn/kong

The world's first agentic reverse engineer.

Repository: https://github.com/amruth-sn/kong
Canonical: https://ross.abutalabs.com/products/amruth-sn-kong
Homepage: https://docs.kong.fyi
Language: Python
License: Apache-2.0
License Family: permissive
Topics: agentic-ai, ghidra, reverse-engineering
Last push: 2026-04-03T02:39:01+00:00

## Health v2 (maintenance only)
Score: 63/100 (v2, computed 2026-09-03T02:39:23.370411+00:00)
- activity 75, release rhythm 75, longevity 13
- inputs: {"age_days": 191, "days_push": 153, "days_rel": 169, "gap_med": 1, "n_releases_24m": 4}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1087, forks 125 (observed 2026-08-28T04:03:32.327565+00:00)

## What it is
Kong is an LLM-orchestrated reverse engineering CLI that plugs AI into Ghidra's analysis engine to recover function names, types, and structures from stripped binaries. It runs a fully autonomous pipeline—triage, call-graph-ordered analysis, agentic deobfuscation, and semantic synthesis—writing results back into Ghidra's program database.

## Use cases
- recover function names and types from a stripped binary
- automatically deobfuscate control-flow flattening and opaque predicates
- analyze malware samples without manual symbol recovery
- rename FUN_00401a30-style functions across a whole binary
- prepare annotated binaries for CTF or security research
- export recovered symbols and structs as JSON after analysis

## When to choose
- you need to reverse engineer stripped or obfuscated binaries faster than manual analysis
- you already use Ghidra and want LLM-assisted renaming, retyping, and struct recovery
- you want a one-command autonomous pipeline instead of babysitting an LLM on raw decompiler output
- you work in malware analysis, CTFs, or vulnerability research on binaries with hundreds of functions

## When to avoid
- you need a GUI-driven interactive reversing workflow rather than a batch CLI pipeline
- you cannot share binary contents with cloud LLM providers like Anthropic or OpenAI
- you don't want to install Ghidra, JDK 21+, and Python 3.11+ dependencies
- your task is dynamic analysis or debugging rather than static binary annotation

## Facets
- artifact type: cli-tool
- maturity: active
- function: reverse-engineering, agent-framework, llm-inference, security, cli
- domain: reverse-engineering, security, artificial-intelligence, developer-tools
- platform: cli, python, cross-platform
- tags: ghidra, binary-analysis, deobfuscation, decompiler, malware-analysis, stripped-binaries, agentic-ai, ctf

## Member repositories
- amruth-sn/kong (main) score 63

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:32.327565+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T06:49:26.072795+00:00, confidence not recorded.
  - readme: https://github.com/amruth-sn/kong (fetched 2026-08-28T04:03:32.327565+00:00, sha f712fde67168)
  - homepage: https://docs.kong.fyi (fetched 2026-08-29T12:52:16.692089+00:00, sha 7fe38ead1d99)
  - site_page: https://docs.kong.fyi/getting-started/what-is-kong (fetched 2026-08-29T12:52:16.701158+00:00, sha 8131b93aadb5)
  - site_page: https://docs.kong.fyi/getting-started/quickstart (fetched 2026-08-29T12:52:16.703321+00:00, sha 16fa17e4c63a)
  - site_page: https://docs.kong.fyi/getting-started/installation (fetched 2026-08-29T12:52:16.705171+00:00, sha 16b608b4d540)
  - site_page: https://docs.kong.fyi/getting-started/setup-wizard (fetched 2026-08-29T12:52:16.706808+00:00, sha 538a115bdb5e)
- Data as of 2026-08-30T08:39:29.467469+00:00.