← ClaudeAtlas

heroku-cli-investigatorlisted

Step-by-step guidance for using the Heroku CLI (~>10.17.0) to inspect dynos, logs, releases, builds, CI runs, and Postgres health during operational incidents without creating or modifying resources. Trigger the skill when investigating failed deploys, stuck dynos, long-running queries, or anomalous build/CI results on a particular Heroku app.
dgmdan/heroku-cli-investigator-skill · ★ 0 · DevOps & Infrastructure · score 57
Install: claude install-skill dgmdan/heroku-cli-investigator-skill
# Heroku CLI Investigator ## Overview Use the Heroku CLI as a forensic console that reports what is currently happening inside an app. Favor commands that only read state—release history, dyno status, realtime logs, builds, CI pipelines, and Postgres diagnostics—and treat this skill as the checklist you run before asking for write access. Look up exact syntax in `references/heroku-cli-commands.md` when you need the flags. ## Preparation & guardrails - Verify the CLI version (`heroku --version`) and keep `~> 10.17.0` on your path so the documented flags remain valid. Authenticate once with `heroku login`, confirm the identity with `heroku whoami`, and export the target app via `HEROKU_APP` or add `--app <name>` to every command. - Install the builds helper if it is not present: `heroku plugins:install @heroku-cli/heroku-builds`. Check `heroku plugins` to confirm it appears. - Avoid running commands that create resources (`heroku apps:create`, `heroku addons:create`, `heroku pg:reset`, etc.) or alter infrastructure; ask for explicit permission before doing anything that writes. Immediately abort if a command prompts to change a pipeline, rotate credentials, or delete data. ## Runtime & dyno inspection - List the live dynos and their states with `heroku ps --app <app>`. Use `heroku logs --app <app>` to stream output and filter by process (`--ps`, `--source`) when investigating a specific worker or web dyno. The same log stream shows real-time output from `heroku run:detached`