[nivSwisa1]

At low effort the model kept the verdict in thinking, so nothing showed or was reported until the hold blocked an edit; the skill now records it with one printf call and a PostToolUse hook shows it on screen (6/6 real sessions at low/medium/max, 98 hook tests).

🤖 Generated with Claude Code

[orca-security-us]

Orca Security Scan Summary

Status	Check	Issues by priority
Passed	Infrastructure as Code	0   0   0   0	View in Orca
Passed	OSS Licenses	0   0   0   0	View in Orca
Passed	SAST	0   0   0   0	View in Orca
Passed	Secrets	0   0   0   0	View in Orca
Passed	Vulnerabilities	0   0   0   0	View in Orca

[upwind-code-us]

Upwind IaC Scan - ✅ Passed

0 newly introduced misconfigurations · 0 resolved · 0 total in this PR vs main

View full analysis in Upwind Console →

Scan completed in 12s

Scan history (2 scans)

Commit	Scanned at	New	Resolved	Net
a184c44	2026-11-10 09:50 UTC	0	0	0
53babf8 <	2026-11-10 09:53 UTC	0	0	0

Last scanned: 53babf8 · 2026-11-10 09:53 UTC

[upwind-code-us]

Upwind Secrets Scan

0 newly introduced secrets · 0 resolved · 3 total in the files this PR changed vs main

Total breakdown: 🔶 3 PII

A secret that has been pushed remains in the repository's history even after the line is removed.

Scan completed in 12s.

---

Scan history (2 scans)

Commit	Scanned at	New	Resolved	Net
a184c44	2026-11-10 09:50 UTC	0	0	0
53babf8 <	2026-11-10 09:53 UTC	0	0	0

Last scanned: 53babf8 · 2026-11-10 09:53 UTC

[upwind-code-us]

Upwind SAST Scan - ✅ Passed

✅ No SAST findings detected

View full analysis in Upwind Console →

Scan completed in 23s

Scan history (2 scans)

Commit	Scanned at	New	Resolved	Net
a184c44	2026-11-10 09:50 UTC	0	0	0
53babf8 <	2026-11-10 09:53 UTC	0	0	0

Last scanned: 53babf8 · 2026-11-10 09:53 UTC

[Copilot]

🟡 Changes recommended

The verdict extraction regex truncates valid plans containing apostrophes; it needs to read tool_response.stdout instead, with fixtures covering escaped apostrophes.

1 open issue

• Verdict parsing truncates plans containing apostrophes · New

What changed in this PR

This PR makes Agentic Advisor effort verdicts visible at every effort level by requiring a dedicated printf call and displaying the verdict through a PostToolUse hook.

Changes:

• Updates skill instructions and edit-hold guidance to require the verdict-printing call.
• Adds hook handling and tests for verdict display, edit unblocking, and reporting.
• Bumps the plugin to version 1.5.2 and documents the fix.

File	Description
tests/​agentic-advisor/​hooks_test.sh	Adds verdict-display and printf-based reporting tests.
plugins/​agentic-advisor/​skills/​agentic-advisor/​SKILL.md	Requires a dedicated verdict-printing Bash call.
plugins/​agentic-advisor/​hooks/​hooks.json	Registers the Bash PostToolUse hook.
plugins/​agentic-advisor/​hooks/​agentic-advisor-trigger.sh	Displays verdicts and updates edit-hold guidance.
plugins/​agentic-advisor/​CHANGELOG.md	Documents the verdict-visibility fix.
plugins/​agentic-advisor/​.claude-plugin/​plugin.json	Updates the plugin version to 1.5.2.

_{Review effort: Balanced}

[MishaKav]

LGTM

[Copilot]

🔵 Needs a closer look

Shell-source parsing truncates verdicts containing apostrophes, and supported manual invocations with the feature disabled lack a visible assistant-text fallback.

1 open issue

• Verdict parsing truncates plans containing apostrophes

1 previously missed issue

In code that hasn't changed since last review

Add visible fallback for manual invocation when disabled

plugins/​agentic-advisor/​skills/​agentic-advisor/​SKILL.md:31

The skill still explicitly supports invocation by name while LINEARB_AGENTIC_ADVISOR_ENABLED is off (see the Off switch section and README.md:91), but the trigger exits before PostToolUse handling in that configuration. Requiring only printf therefore leaves these explicitly requested verdicts inside the collapsed Bash output, with no display message. Preserve the off switch, but require a visible assistant-text fallback for this supported manual-invocation path.

_{Review effort: Balanced}