Dashboard › cli › Distillation
a6ea1aed-001f-43c3-81ae-d6833a9efa32["lore_tm_v1_fD4GYKZg39yPuX7zSzN0UkOwx6wW2XRs38a4hnaM9e8","lore_tm_v1_8j5tIytLADDqgyv_5KXWDUC6B1VGhC094BRNmw49nMk","lore_tm_v1_rFDxbc1a-rhLV2vuDnr1LYTxUI7yWXbcISgWjLFCsQE","lore_tm_v1_vyJEL49AbnyouCmvAVt6jYoVicbQtvWd7M2ez9sKqDk","lore_tm_v1_CVana0muiZCSu71_kRrTHibKFo7NYGF1wxm9oJGPaj4","lore_tm_v1_z07MZVNfnpmoLjB9wiRY-MQ2Qn7ZK3Xziq87hdCjSk4"]
.github/workflows/pr-risk.yml defines PR Risk Experiment on pull_request events opened, synchronize, reopened, and ready_for_review; grants contents: read, pull-requests: write, and issues: write; uses concurrency group pr-risk-${{ github.event.pull_request.number }} with cancel-in-progress: true; skips draft PRs; runs getsentry/pr-risk-action@v0 with result-path: risk-pr-result.json, skip-reviews: "true", and apply-label: ${{ github.event.pull_request.head.repo.full_name == github.repository }}; then uploads risk-pr-result.json as artifact risk-pr-result via actions/upload-artifact@v4.getsentry/pr-risk-action@v0 documentation found the action scores PRs as low, medium, or high, writes GITHUB_STEP_SUMMARY and risk-pr-result.json, exposes logistic/rule/final risk outputs, and synchronizes exactly one of risk: low, risk: medium, or risk: high; its documented safety model is to avoid checking out or executing pull-request code and instead read metadata/diffs through the GitHub API.getsentry/pr-risk-action@v0/action.yml found inputs repo, pr-number, profile, history, model, result-path (default risk-pr-result.json), skip-reviews (default "true"), apply-label (default "true"), label-prefix (default "risk: "), and python-version (default "3.11"); outputs include risk-label, final-risk-label, final-risk-policy, logistic-risk-label, rule-risk-label, logistic-probability, logistic-percentile-repo, rule-percentile-repo, and github-label.getsentry/pr-risk-action@v0/action.yml is a composite action that runs actions/setup-python@v5, installs itself with python -m pip install "$GITHUB_ACTION_PATH", derives the default profile as ${INPUT_REPO//\//__}, defaults history/model to $GITHUB_ACTION_PATH/profiles/$profile/pr-history.jsonl.gz and $GITHUB_ACTION_PATH/profiles/$profile/model.json, fails if either file is missing, and invokes risk-pr score-pr with --repo, --pr, --history, --model, --out, and --summary-file; it conditionally appends --skip-reviews and --label-pr --label-prefix.score_pull_request() fetches the current PR through GitHubClient, loads bundled historical rows, removes any historical row for the target PR, deduplicates by (repo, number), requires at least one historical row plus the current PR, builds features, applies the serialized logistic model, and produces a Markdown summary.apply_pr_risk_label() selects final_risk_label, then logistic_risk_label, then deterministic risk_label; validates against low, medium, and high; upserts the selected label; removes other labels sharing the configured risk prefix; and adds the selected label. cmd_score_pr() records label failures, prints a warning, emits output/summary data, and then re-raises the GitHubError..gitignore ignores apps/cli-docs/src/content/docs/commands/ and apps/cli-docs/public/.well-known/skills/index.json; packages/cli/.gitignore ignores src/generated/, including packages/cli/src/generated/skill-content.ts.packages/cli/plugins/sentry-cli/skills/sentry-cli/references/agent-conversation.md found generated skill documentation version 0.45.0-dev.0 renders list usage as sentry agent-conversation list [<org>] and view usage as sentry agent-conversation view [<org>/]<conversation-id>.sentry agent-conversation view conv-123; explicit organization with sentry agent-conversation view my-org/conv-123; JSON output with sentry agent-conversation view my-org/conv-123 --json.sentry agent-conversation list; sentry agent-conversation list my-org; sentry agent-conversation list --limit 50 --period 24h; sentry agent-conversation list -q "has:errors"; and sentry agent-conversation list my-org -c next.