Skip to content

feat(cli): Display confidence level in findings output#194

Merged
dcramer merged 3 commits intomainfrom
claude/slack-session-RWLFj
Feb 20, 2026
Merged

feat(cli): Display confidence level in findings output#194
dcramer merged 3 commits intomainfrom
claude/slack-session-RWLFj

Conversation

@dcramer
Copy link
Member

@dcramer dcramer commented Feb 20, 2026

Add confidence level display to findings in CLI output

When triaging findings, severity alone doesn't tell the full story. A high-severity finding with low confidence needs very different treatment than one with high confidence. This surfaces confidence as a color-coded badge (green/yellow/red) after the description in both TTY output and the interactive fix flow, so users can make better triage decisions at a glance.

Badge is omitted entirely when confidence is undefined, keeping output clean for skills that don't report it.

┌── security-review ──────────────────────── [4.2s] ──┐
│ 1 finding: ● 1 high                                 │
├──────────────────────────────────────────────────────┤
│                                                      │
│ ● (high) SQL injection risk in query builder         │
│   src/api/auth.ts:42  +0.8s                          │
│   42 │ const query = buildQuery(userId);              │
│                                                      │
│   User input is interpolated into a SQL query...     │
│                                                      │
│   [high confidence]                                  │
│                                                      │
│   Suggested fix:                                     │
│   - const query = buildQuery(userId);                │
│   + const query = buildQuery(sanitize(userId));      │
│                                                      │
└──────────────────────────────────────────────────────┘

Closes #190

Add confidence badge display to the CLI interactive output, allowing
users to assess finding reliability alongside severity. This helps with
triaging findings since high-severity + low-confidence is very different
from high-severity + high-confidence.

Changes:
- Add formatConfidenceBadge() and formatConfidencePlain() formatters
- Display confidence in TTY mode after severity badge
- Display confidence in fix flow interactive prompt
- Add tests for new confidence display functionality

Closes #190

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

https://claude.ai/code/session_011XdKDgKtk2tuTZ5Y9GsG5i
Copy link

@cursor cursor bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Cursor Bugbot has reviewed your changes and found 1 potential issue.

Bugbot Autofix is OFF. To automatically fix reported issues with Cloud Agents, enable Autofix in the Cursor dashboard.

Confidence is now rendered after the description and verification lines
(separated by a blank line) instead of inline on the title line. This
keeps the title line clean and scannable while still surfacing confidence
as a distinct metadata field.

Updates the reporters spec examples and formatting functions table to
document the new placement and the two confidence formatter functions.

Closes #190

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
@dcramer dcramer changed the title Add confidence level display to findings in terminal output feat(cli): Display confidence level in findings output Feb 20, 2026
Dead code flagged by Cursor Bugbot. The CI rendering path inlines
confidence formatting directly, so this function was never called.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
@dcramer dcramer merged commit 9b02bcc into main Feb 20, 2026
12 checks passed
@dcramer dcramer deleted the claude/slack-session-RWLFj branch February 20, 2026 05:46
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Render confidence alongside severity in CLI interactive output

2 participants

Comments