fix: harden child-process argv and whisper path passing

[undivisible]

What

Closes leftover Jules injection/flag-confusion gaps without changing intended sh -c skill/test semantics.

• Telegram faster-whisper: static python3 -c body; audio path is sys.argv[1], never interpolated (Jules 🔒 Fix Code Injection Vulnerability in Telegram Voice Transcription #55).
• Dynamic tools: -- before python/bash script path; scrub secret env vars (Jules 🔒 Fix command injection vulnerability in dynamic tool execution #71).
• Vibemania: flags first, then -- + goal; env scrub (Jules 🔒 Fix command injection and secret leak in vibemania tool #61).
• Local message bridge: log bind/serve errors instead of unwrapping.

Tests

Unit coverage for whisper argv construction and dynamic-tool -- placement. Full --all-features suite did not finish in the orb (RocksDB compile).

Review-only. Process-spawn / security boundary — do not auto-merge.

---

Note

Medium Risk
Touches security-sensitive process spawning and untrusted path/argument handling; changes are defensive but warrant careful review of argv ordering and env scrubbing on each spawn site.

Overview
Hardens several child-process spawn paths against injection and flag confusion without changing shell-based skill semantics elsewhere.

Telegram voice transcription now runs a fixed python3 -c Whisper script and passes the temp audio file only as sys.argv[1] (not embedded in the script), with kill_on_drop and child_proc::scrub on the Python process. A unit test locks in that behavior.

Dynamic tools (python3 / bash) insert -- before the script path so paths starting with - are not parsed as flags; execution is centralized via dynamic_tool_argv, with the same scrub/kill-on-drop treatment. Vibemania applies the same pattern: flags first, then -- plus the user goal.

The local Telegram message bridge logs bind/serve failures instead of panicking on unwrap.

New tests cover Whisper argv construction and dynamic-tool -- placement.

^{Reviewed by Cursor Bugbot for commit a63d4f0. Configure here.}

[capy-ai]

Capy couldn't review this pull request because Max's workspace is out of credits, add credits or enable auto-reload to resume automatic reviews.

[cursor]

Bugbot couldn't run - usage limit reached

Bugbot is counted against Cursor usage for this user or team, and this run hit a usage or spend limit.

A user or team admin can review and increase usage limits in the Cursor dashboard.

(requestId: serverGenReqId_826a458d-b09f-4a16-bad3-3a9ab68f7a7c)

[mergify]

Tick the box to add this pull request to the merge queue (same as @mergifyio queue).

• Queue this pull request

[cursor]

Bugbot couldn't run - usage limit reached

Bugbot is counted against Cursor usage for this user or team, and this run hit a usage or spend limit.

A user or team admin can review and increase usage limits in the Cursor dashboard.

(requestId: serverGenReqId_5aea07cd-8297-47c5-b752-dd588af5b193)