UAT: Shell danger detection missing several spec-required patterns — fork bomb with spaces, rm -r, chmod 777, fdisk, parted, kill -9, sudo #1428

New issue

Open

opened 2026-04-02 17:50:16 +00:00 by freemo · 1 comment

freemo commented

2026-04-02 17:50:16 +00:00

Owner

Metadata

Branch: bugfix/m8-shell-danger-detection-missing-patterns
Commit Message: fix(tui): add missing spec-required danger patterns to shell_exec looks_dangerous
Milestone: v3.7.0
Parent Epic: #341

Bug Report

What was tested: Shell mode dangerous command detection (src/cleveragents/tui/input/shell_exec.py)

Expected behavior (from spec):
The specification (Section "Shell Danger Detection") lists these patterns that should be detected:

Pattern	Risk Level
`rm -rf` / `rm -r`	High
`chmod 777`	Medium
`> /dev/sda` / `dd if=`	High
`:(){ :	:& };:` (fork bomb)
`mkfs` / `fdisk` / `parted`	High
`kill -9` / `killall`	Medium
`sudo` / `su`	Low

Actual behavior:
The implementation in src/cleveragents/tui/input/shell_exec.py only detects these patterns:

patterns = (
    "rm -rf /",
    "git push --force",
    "mkfs.",
    "dd if=",
    ":(){:|:&};:",
)

The following spec-required patterns are not detected:

rm -r (without /) — rm -r /tmp is not flagged as dangerous
:(){ :|:& };: (fork bomb with spaces) — the implementation uses :(){:|:&};: (no spaces), so the standard fork bomb pattern with spaces is not detected
chmod 777 — not detected
> /dev/sda — not detected
fdisk — not detected
parted — not detected
kill -9 / killall — not detected
sudo / su — not detected

Steps to reproduce:

from cleveragents.tui.input.shell_exec import looks_dangerous

print(looks_dangerous("rm -r /tmp"))          # Returns False (should be True)
print(looks_dangerous(":(){ :|:& };:"))        # Returns False (should be True)
print(looks_dangerous("chmod 777 /var/www"))   # Returns False (should be True)
print(looks_dangerous("fdisk /dev/sda"))       # Returns False (should be True)
print(looks_dangerous("kill -9 1234"))         # Returns False (should be True)
print(looks_dangerous("sudo rm -rf /"))        # Returns True (correct, but only because rm -rf / matches)

Code location: src/cleveragents/tui/input/shell_exec.py function looks_dangerous()

Note: The spec states "Danger detection is controlled by the shell.warn_dangerous setting (default: true). The detection is advisory only — it never prevents command execution." So this is about warning display, not blocking.

Severity: Medium — users may not receive warnings for potentially destructive commands.

Subtasks

Add rm -r (without trailing /) to the danger patterns in looks_dangerous()
Add :(){ :|:& };: (fork bomb with spaces) alongside the existing no-spaces variant
Add chmod 777 to the danger patterns
Add > /dev/sda to the danger patterns
Add fdisk to the danger patterns
Add parted to the danger patterns
Add kill -9 and killall to the danger patterns
Add sudo and su to the danger patterns
Tests (Behave): Add/update scenarios in features/ covering each newly added pattern, asserting looks_dangerous() returns True for each
Tests (Behave): Add regression scenarios asserting previously-passing patterns still return True
Tests (Robot): Add/update integration smoke test verifying the TUI shell danger warning is displayed for the new patterns
Run nox -e typecheck — confirm 0 errors
Verify coverage >= 97% via nox -s coverage_report
Run nox (all default sessions), fix any errors

Definition of Done

All subtasks above are completed and checked off.
looks_dangerous() correctly returns True for all patterns listed in the spec: rm -r, :(){ :|:& };: (with spaces), chmod 777, > /dev/sda, fdisk, parted, kill -9, killall, sudo, su.
No previously-passing danger patterns are broken by the change.
A Git commit is created where the first line of the commit message matches the Commit Message in Metadata exactly, followed by a blank line, then additional lines providing relevant details about the implementation.
The commit is pushed to the remote on the branch matching the Branch in Metadata exactly.
The commit is submitted as a pull request to master, reviewed, and merged before this issue is marked done.
All nox stages pass
Coverage >= 97%

## Metadata - **Branch**: `bugfix/m8-shell-danger-detection-missing-patterns` - **Commit Message**: `fix(tui): add missing spec-required danger patterns to shell_exec looks_dangerous` - **Milestone**: v3.7.0 - **Parent Epic**: #341 ## Bug Report **What was tested:** Shell mode dangerous command detection (`src/cleveragents/tui/input/shell_exec.py`) **Expected behavior (from spec):** The specification (Section "Shell Danger Detection") lists these patterns that should be detected: | Pattern | Risk Level | |---------|-----------| | `rm -rf` / `rm -r` | High | | `chmod 777` | Medium | | `> /dev/sda` / `dd if=` | High | | `:(){ :|:& };:` (fork bomb) | High | | `mkfs` / `fdisk` / `parted` | High | | `kill -9` / `killall` | Medium | | `sudo` / `su` | Low | **Actual behavior:** The implementation in `src/cleveragents/tui/input/shell_exec.py` only detects these patterns: ```python patterns = ( "rm -rf /", "git push --force", "mkfs.", "dd if=", ":(){:|:&};:", ) ``` The following spec-required patterns are **not detected**: 1. `rm -r` (without `/`) — `rm -r /tmp` is not flagged as dangerous 2. `:(){ :|:& };:` (fork bomb with spaces) — the implementation uses `:(){:|:&};:` (no spaces), so the standard fork bomb pattern with spaces is not detected 3. `chmod 777` — not detected 4. `> /dev/sda` — not detected 5. `fdisk` — not detected 6. `parted` — not detected 7. `kill -9` / `killall` — not detected 8. `sudo` / `su` — not detected **Steps to reproduce:** ```python from cleveragents.tui.input.shell_exec import looks_dangerous print(looks_dangerous("rm -r /tmp")) # Returns False (should be True) print(looks_dangerous(":(){ :|:& };:")) # Returns False (should be True) print(looks_dangerous("chmod 777 /var/www")) # Returns False (should be True) print(looks_dangerous("fdisk /dev/sda")) # Returns False (should be True) print(looks_dangerous("kill -9 1234")) # Returns False (should be True) print(looks_dangerous("sudo rm -rf /")) # Returns True (correct, but only because rm -rf / matches) ``` **Code location:** `src/cleveragents/tui/input/shell_exec.py` function `looks_dangerous()` **Note:** The spec states "Danger detection is controlled by the `shell.warn_dangerous` setting (default: `true`). The detection is advisory only — it never prevents command execution." So this is about warning display, not blocking. **Severity:** Medium — users may not receive warnings for potentially destructive commands. ## Subtasks - [ ] Add `rm -r` (without trailing `/`) to the danger patterns in `looks_dangerous()` - [ ] Add `:(){ :|:& };:` (fork bomb with spaces) alongside the existing no-spaces variant - [ ] Add `chmod 777` to the danger patterns - [ ] Add `> /dev/sda` to the danger patterns - [ ] Add `fdisk` to the danger patterns - [ ] Add `parted` to the danger patterns - [ ] Add `kill -9` and `killall` to the danger patterns - [ ] Add `sudo` and `su` to the danger patterns - [ ] Tests (Behave): Add/update scenarios in `features/` covering each newly added pattern, asserting `looks_dangerous()` returns `True` for each - [ ] Tests (Behave): Add regression scenarios asserting previously-passing patterns still return `True` - [ ] Tests (Robot): Add/update integration smoke test verifying the TUI shell danger warning is displayed for the new patterns - [ ] Run `nox -e typecheck` — confirm 0 errors - [ ] Verify coverage >= 97% via `nox -s coverage_report` - [ ] Run `nox` (all default sessions), fix any errors ## Definition of Done - [ ] All subtasks above are completed and checked off. - [ ] `looks_dangerous()` correctly returns `True` for all patterns listed in the spec: `rm -r`, `:(){ :|:& };:` (with spaces), `chmod 777`, `> /dev/sda`, `fdisk`, `parted`, `kill -9`, `killall`, `sudo`, `su`. - [ ] No previously-passing danger patterns are broken by the change. - [ ] A Git commit is created where the **first line** of the commit message matches the Commit Message in Metadata exactly, followed by a blank line, then additional lines providing relevant details about the implementation. - [ ] The commit is pushed to the remote on the branch matching the **Branch** in Metadata exactly. - [ ] The commit is submitted as a **pull request** to `master`, reviewed, and **merged** before this issue is marked done. - [ ] All nox stages pass - [ ] Coverage >= 97%

freemo added the

labels

2026-04-02 17:50:39 +00:00

freemo commented

2026-04-02 17:51:07 +00:00

Author

Owner

Dependency note: This issue is a child of the TUI Epic #341 (feat(ui): add TUI/Web interface). Per the project hierarchy rules, this issue blocks #341 — the Epic cannot be considered complete until this bug is resolved.

This issue is independent of any currently-open issue and does not block any other in-progress work. It can be picked up in parallel with other TUI work.

**Dependency note:** This issue is a child of the TUI Epic #341 (`feat(ui): add TUI/Web interface`). Per the project hierarchy rules, this issue **blocks** #341 — the Epic cannot be considered complete until this bug is resolved. This issue is **independent** of any currently-open issue and does not block any other in-progress work. It can be picked up in parallel with other TUI work.

freemo referenced this issue

2026-04-02 17:51:07 +00:00

feat(ui): add TUI/Web interface #341

freemo referenced this issue

2026-04-02 18:22:31 +00:00

[UAT-SUPERVISOR] Progress Report — Cycle 11 — v3.7.0 TUI #1453

freemo referenced this issue

2026-04-02 18:27:35 +00:00

[UAT-SUPERVISOR] Progress Report — Cycle 21 — v3.7.0 TUI #1454

freemo referenced this issue

2026-04-02 18:32:39 +00:00

[UAT-SUPERVISOR] Progress Report — Cycle 31 — v3.7.0 TUI #1455