Deepline Feedback
SkillDev toolsSend feedback or bug reports to the Deepline team. Use when the user asks to report feedback or share a session, and proactively when a Deepline Play is disproportionately slow, appears stuck, or an ordinary outcome requires avoidable product steps.
Available today. Use it from your connected AI after setup.
No other account needed.
Connect ahel once, and every AI you use reads what you have installed.
Then ask your AI: use the Deepline Feedback skill
What this skill tells your AI
The instructions your AI receives, as published by getaero-io/gtm-eng-skills in skills/deepline-feedback/SKILL.md and read by ahel’s review.
Quick Start
npm install -g deepline
# Fallback for secure sandboxes: mkdir -p "$HOME/.local" && npm config set prefix "$HOME/.local" && export PATH="$HOME/.local/bin:$PATH" && npm install -g deepline --registry https://code.deepline.com/api/v2/npm/
deepline auth register --wait auto
deepline auth wait --timeout 120 # completes Cowork/browser approval; no-op if already connected
deepline auth status
deepline -h
CLI resolution
Run deepline when it is available. If the shell reports that command is missing, use <workspace-root>/.deepline/runtime/bin/deepline (or the npm-created .cmd shim on Windows). If neither exists, follow https://code.deepline.com/INSTALL.md to set up Deepline.
Send feedback or a bug report to the Deepline team.
Steps
-
Get feedback text. Use the argument if provided (e.g.
/deepline-feedback the waterfall broke). Otherwise ask the user. -
Confirm. Use AskUserQuestion with a question like:
This report will include:
- Your feedback: {feedback text}
- Environment info (auto-collected)
- Current session transcript
Send this feedback?
Options: "Send it" / "Cancel".
-
If confirmed, send the feedback text first. Because the user asked for this report, pass
--requestedso Deepline can distinguish it from an autonomous agent report:deepline feedback send "{feedback text}" --requested --jsonFor reports the agent initiates autonomously, omit
--requested. -
Send the session transcript. Try the normal Claude transcript location first:
deepline sessions send --current-session --jsonIf that reports no session files and
~/mnt/.claude/projectsexists, the run is likely in Cowork. Bridge the mounted transcript directory, then retry. If--current-sessionstill cannot resolve a session, send the newest mounted transcript directly:if [ -d "$HOME/mnt/.claude/projects" ]; then mkdir -p "$HOME/.claude" ln -sfn "$HOME/mnt/.claude/projects" "$HOME/.claude/projects" 2>/dev/null || true deepline sessions send --current-session --json || deepline sessions send --file "$(ls -t "$HOME"/mnt/.claude/projects/*/*.jsonl | head -1)" --json fiUse the plural
sessions sendcommand, not the old singular session form. -
Tell the user it was sent. If cancelled, do nothing.
Signals
- GitHub stars
- 58
- Forks
- 11
- Last commit
- Sep 2026
Advanced
- Catalog kind
- skill
- Gateway key
deepline-feedback- Source
- github.com/getaero-io/gtm-eng-skills