Repository navigation
Update README.md - #7
Conversation
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 2c7b2c91d5
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| QCLI_PINNED_MODEL=model-id | ||
|
|
||
| # Force deterministic mode without passing --no-ai | ||
| QCLI_NO_AI=1 |
There was a problem hiding this comment.
Do not tell users
.env disables AI
If QCLI_NO_AI=1 is set only in the suggested .env file, it does not disable AI: the provider resolver reads that file, but no_ai_mode() checks only the process environment. With ABLITERATION_KEY also configured, planning can still make provider requests. Use a shell environment variable or --no-ai for the offline path.
AGENTS.md reference: AGENTS.md:L15-L16
Useful? React with 👍 / 👎.
| ### Exploitation | ||
|
|
||
| ```bash | ||
| python quantum_cli.py auto -t 10.0.0.42 # validator-gated auto-exploit; --force skips the gate |
There was a problem hiding this comment.
Avoid presenting
auto -t as a single-host exploit
When a user runs this example, main() handles any --target before dispatching the selected command: a bare IP is expanded to its /24 and sent to AutonomousOverlord, so auto_exploit() is never called. This runs an autonomous camera-first engagement over a broader scope instead of the documented targeted exploit.
AGENTS.md reference: AGENTS.md:L31-L31
Useful? React with 👍 / 👎.
| | `--cycles <N>` | Autonomous cycle cap. `0` is infinite. | | ||
| | `--aggression balanced\|aggressive\|patient` | Autonomous posture at the validator gate. | | ||
| | `--model <id>` | Pin the AI model. | | ||
| | `--kill` | Arm the kill-switch against a live C2 via `POST /kill`. | |
There was a problem hiding this comment.
Make
--kill invoke the kill-switch before documenting it
If an operator invokes the advertised --kill option by itself, main() never reads a.kill and falls through to the no-command autonomous path; with serve, it starts the C2 without arming the switch. The documented flag therefore does not kill implants and can start an unintended engagement instead.
AGENTS.md reference: AGENTS.md:L31-L31
Useful? React with 👍 / 👎.
|
|
||
| | Symptom | Fix | | ||
| | --- | --- | | ||
| | Port 8443 already bound | `python quantum_cli.py serve --port 8444`, or kill the stale process | |
There was a problem hiding this comment.
Use a port override that the C2 server honors
When port 8443 is already bound, the suggested serve --port 8444 does not resolve the conflict: the serve handler calls run_c2() without passing the parsed port, so it still binds the default 8443.
AGENTS.md reference: AGENTS.md:L31-L31
Useful? React with 👍 / 👎.
| python quantum_cli.py listener deploy # allocate a listener on 9100-9999 | ||
| python quantum_cli.py listener active # list active listeners | ||
| python quantum_cli.py listener hits # show callback hits | ||
| python quantum_cli.py listener stopall # tear down every listener |
There was a problem hiding this comment.
Use the listener command's query option in these examples
As written, each listener example passes an extra positional argument and argparse rejects it; the CLI takes the listener action through -q/--query instead. Use that option for deploy, active, hits, and stopall so these commands reach their handlers.
AGENTS.md reference: AGENTS.md:L31-L31
Useful? React with 👍 / 👎.
| python quantum_cli.py wipe # footprint wipe, per host OS | ||
| python quantum_cli.py seal # seal all loot (GRAYFISH) | ||
| python quantum_cli.py unseal # unseal all loot | ||
| python quantum_cli.py harvest # pull staged artifacts |
There was a problem hiding this comment.
Remove or implement the unsupported
harvest command
Running this documented command exits with an invalid-choice error: harvest is absent from the CLI's top-level command choices and has no direct command handler.
AGENTS.md reference: AGENTS.md:L31-L31
Useful? React with 👍 / 👎.
No description provided.