Vtai Agent

Install

Works with Claude Code, Codex, VS Code, Cursor, Windsurf, Claude Desktop. Read the repository before you give any tool access to your data.

npx agentmarket install connector/velsrocky-vtai-agent

About

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this server can do and what it asks for, from its tool list, install config and code. Checked Oct 25, 2026.

Tools

3 read · 1 write · 1 delete · 4 not yet classified

Secrets

None asked for

Install scripts

None found

Runs as

Python package

Tools (9)

Read from the source code; the running server may differ. Descriptions are the server’s own words; hidden characters are shown as [U+…].

• auto_mode

  Autonomous loop: dispatch the goal to a coding CLI (opencode by default), review the reply against the goal, and iterate with corrective prompts until the answer satisfies the goal or max_iterations is reached. Use this for the '@vt-ai_automode_on' keyword. Args: goal: The full goal or question to satisfy. working_dir: Project directory the CLI runs in (must be writable). cli: opencode | claude | pi max_iterations: Max review/reprompt rounds (1-10). timeout: Per-iteration seconds before the CLI is killed. dry_run: Preview only (defaults to config dry_run_default).

• backup_sync

  Incremental directory backup via rsync. Reports the transfer plan and can verify with a checksum pass. Deletion of extraneous files is opt-in and never happens during a dry run. Args: source: Directory to back up. destination: Backup destination (must be in writable_roots). dry_run: Preview only (defaults to config dry_run_default). delete_extraneous: Delete destination files not in source (dangerous). compress: Use rsync -z. verify: Re-run with checksums after transfer.

• delegate_coding

  Delegate a coding task to opencode, claude, or pi in headless mode. Snapshots the repo, captures the diff, and rolls back if an optional verify command fails. The working directory must be a git repo. Args: cli: opencode | claude | pi prompt: The coding task to perform. working_dir: Project directory (must be a git repo, in writable_roots). dry_run: Snapshot and plan only (defaults to config dry_run_default). verify_command: Allowlisted command that must pass afterward, e.g. "pytest". auto_rollback: Restore the snapshot if verification fails. require_approval: Keep…

• disk_audit

  Read-only scan for reclaimable disk space: largest/oldest files plus cache directory sizes. Deletes nothing. Args: path: Root to scan. min_size_mb: Only report files at least this large. older_than_days: Only report files older than this. limit: Max files to report. include_caches: Also measure known cache directories.

• inspect_application

  Analyze a local application with REA (reverse engineering) under full guardrails. The target is deny-glob checked, copied into a quarantine dir, analyzed there (REA capability must be enabled in config), and the returned evidence carries an audit digest. The original target is never touched. Args: goal: What you want to learn from the target app. target: Path to the app (directory, ASAR, or binary). judge: Verify the evidence answers the goal with the judge model. keep_copy: Keep the quarantine copy after the run for audit. timeout: Seconds before REA is killed (config def…

• organize_files

  Content-aware cleanup: sort a directory's files into Documents/Images/ Videos/Audio/Archives/Installers/Code/Data/Fonts/Models. Collisions are auto-renamed. Everything stays inside the target directory. Args: directory: Directory to tidy (default ~/Downloads). dry_run: Plan only, move nothing (defaults to config dry_run_default). recursive: Descend into subdirectories. confirm_mime: Use `file` mime type for unknown extensions.

• run_shell

  Run an allowlisted shell command. Protected paths and non-allowlisted binaries are rejected. Use dry_run=true to preview without executing. Args: command: Shell command; argv[0] must be in the allowlist. dry_run: Preview only (defaults to config dry_run_default). timeout: Seconds before the command is killed (1-300).

• system_info

  Snapshot this machine: CPU, memory, disks, load, and AMD GPU state. Read-only and safe to call before deciding how to act. Args: include_gpu: Probe rocm-smi/vulkaninfo for GPU state (default true).

• transcode_media

  Batch transcode videos with AMD GPU (VAAPI) acceleration. Output goes to a 'transcoded' subfolder; originals are never modified. Args: directory: Directory containing videos. dry_run: Report jobs without encoding. codec: h264_vaapi | hevc_vaapi | av1_vaapi | libx264 (config default). crf: Quality for software codecs, 0-51 (lower = better). recursive: Descend into subdirectories. overwrite: Re-encode even if output exists.

Signals

Stars

0

Forks

0

Open issues

0

Weekly downloads

415

Last commit

Oct 22, 2026 (7 days ago)

Last release

Oct 22, 2026 (7 days ago)

Last package publish

Oct 22, 2026 (7 days ago)

From GitHub (fetched Oct 24, 2026) and the package registry.

Details

Registry name

io.github.velsrocky/vtai-agent

Version

0.7.0

Auth

No auth

Packages

pypi vtai-agent

License

MIT

Repository

github.com/velsrocky/vtai-agent

First seen

Oct 24, 2026

Last checked

Oct 24, 2026

Source

Official MCP Registrygithub-graphql