| name | linear-propose-followups |
| description | Draft post-deploy follow-up issues from a list of observations, then apply
human-gate review before creating anything.
Uses the propose-followups verb via linear.mjs. Every issue gets tech-debt +
post-deploy + ai-suggested labels.
Use when: engineer says "propose follow-ups", "log these after the deploy",
"file follow-ups for that deploy", or /linear-propose-followups.
Skip when: the work is a single obvious fix (use /linear-log-bug); or the
observations don't come from a real post-deploy signal.
|
linear-propose-followups
Draft a set of post-deploy follow-up Linear issues with human gate.
Procedure
- Collect observations — one per line — from the engineer. Examples:
- "router cache hit rate dropped from 92% to 78% after deploy"
- "cold start on Lambda regressed to 3s"
- Run the script in proposal mode (default, no
--go):
printf "%s\n" "router caching 92%% -> 78%%" "cold start 1s -> 3s" \
| node "${CLAUDE_PLUGIN_ROOT}/scripts/linear.mjs" propose-followups \
--project "My Project Name" \
--env prod
Prints the draft titles and waits.
- Show the drafts to the engineer. If they say yes, re-run with
--go:
printf "%s\n" "router caching 92%% -> 78%%" "cold start 1s -> 3s" \
| node "${CLAUDE_PLUGIN_ROOT}/scripts/linear.mjs" propose-followups \
--project "My Project Name" \
--env prod --go
- Report the created
TEAM-<id> numbers.
Important
- Batches of 3+ require human gate — never skip step 3.
- PII scan rejects observations that contain DOB / IDs.
- Project name is a substring match; if ambiguous, the script errors and
asks for a more specific query.
- The team key is resolved from
LINEAR_TEAM_KEY env or the plugin's
configured default.