• Fix ragged table output in search and verify commands using stdlib text/tabwriter
• Make truncateDesc rune-safe to prevent UTF-8 corruption from multibyte character splitting
• Add internal/sourceguard test package to guard against accidental non-ASCII runes in source
• Add Claude agent-loop hook in .claude/settings.json to enforce ASCII guard on .go edits

Align table output and fix rune-safe truncation

• Added writeAlignedColumns helper using text/tabwriter for shared table rendering across
 search and verify commands
• Refactored renderSearchResult to build rows and pass through writeAlignedColumns instead of
 direct formatting
• Refactored renderVerifyFailure to align failure reasons using the same writeAlignedColumns
 helper
• Made truncateDesc rune-safe by converting to rune slice before slicing, preventing UTF-8
 corruption from multibyte characters

internal/cli/output.go

Add non-ASCII source hygiene test guard

• New test package that scans all .go files for unsanctioned non-ASCII runes
• Maintains explicit allowlist of intentional glyphs (→, ✓, ✗, —, ·, …, •, §, ≤, –, ⇒) with
 justifications
• Reports violations as file:line:col: U+XXXX 'rune' for easy identification and fixing
• Runs via go test ./... so enforcement is automatic in make check and CI

internal/sourceguard/ascii_test.go

Test helper utilities for source guard

• Helper functions for module root discovery by walking up to go.mod
• Directory skip logic to exclude VCS, tooling, and build directories
• Path formatting utilities for readable violation messages
• Violation formatting as rel:line:col: U+XXXX 'r' for clear error reporting

internal/sourceguard/helpers_test.go

Description

The new tabwriter.NewWriter call passes 6 arguments on a single line, which violates the style
rule requiring one argument per line for calls with 4+ arguments. This reduces readability and makes
future diffs noisier.

Code

internal/cli/output.go[188]

Evidence

PR Compliance ID 782553 requires calls with 4+ arguments to use one argument per line. The newly
added tabwriter.NewWriter(w, 0, 0, 2, ' ', 0) is a 6-argument call on one line.

Rule 782553: Function calls with 4+ arguments must use one argument per line
internal/cli/output.go[188-188]

Agent prompt

The issue below was found during a code review. Follow the provided context and guidance below and implement a solution

## Issue description
A function call with 4+ arguments is written on a single line, but the style guide requires one argument per line.

## Issue Context
`tabwriter.NewWriter(w, 0, 0, 2, ' ', 0)` has 6 arguments.

## Fix Focus Areas
- internal/cli/output.go[188-188]

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools

Description

The new .claude/settings.json hook embeds a very long shell command on a single line, exceeding
the ~120 character guideline. This makes the configuration hard to review and maintain.

Code

.claude/settings.json[9]

Evidence

PR Compliance ID 782552 requires breaking lines exceeding ~120 characters at semantic boundaries.
The added hook command is a single long line containing multiple shell statements and
conditionals.

Rule 782552: Break lines exceeding ~120 characters at semantic boundaries
.claude/settings.json[9-9]

Agent prompt

The issue below was found during a code review. Follow the provided context and guidance below and implement a solution

## Issue description
A single JSON line contains an overlong shell command, exceeding the ~120 char line-length guideline.

## Issue Context
JSON strings can’t be cleanly line-wrapped; the maintainable pattern is to move complex shell logic into a small script file and reference it from the hook.

## Fix Focus Areas
- .claude/settings.json[9-9]

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools

Description

The updated truncation logic appends a Unicode ellipsis … rather than the required .... This
deviates from the compliance requirement for human-output truncation formatting.

Code

internal/cli/output.go[214]

Evidence

PR Compliance ID 783450 requires human-oriented formatting to truncate descriptions to 80 characters
and append ... if truncated. The modified truncateDesc implementation returns ... + "…"
(Unicode ellipsis) instead of ....

Rule 783450: Human output must truncate descriptions to 80 characters and replace newlines with spaces
internal/cli/output.go[206-215]

Agent prompt

The issue below was found during a code review. Follow the provided context and guidance below and implement a solution

## Issue description
Human output truncation is required to append `...`, but the code appends the Unicode ellipsis `…`.

## Issue Context
The truncation budget is `searchDescWidth` (80), so changing the suffix may require adjusting the slice length to keep the final rendered width at 80 characters.

## Fix Focus Areas
- internal/cli/output.go[206-215]

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools