fix(labels): escape single quotes in custom label descriptions

[gvago]

Summary

• Bug: set_custom_labels() in pr_agent/algo/utils.py wraps label descriptions in single quotes without escaping. A description like Don't merge produces the malformed string 'Don't merge', breaking the generated custom_labels_class Enum schema sent to the LLM.
• Fix: Added .replace("'", "\\'") to escape single quotes in descriptions before they are wrapped in single-quote delimiters. This is chained after the existing .replace('\n', '\\n') call on line 959.
• Affects both pr_description.py and pr_generate_labels.py since both call the shared set_custom_labels() utility.

Test plan

• Configure a custom label with a single-quote in its description (e.g., description = "Don't merge this")
• Run /describe or /generate_labels and verify the generated custom_labels_class contains properly escaped strings
• Verify labels without single quotes still work unchanged

🤖 Generated with Claude Code

[qodo-free-for-open-source-projects]

Review Summary by Qodo

Escape single quotes in custom label descriptions

🐞 Bug fix

Walkthroughs

Description

• Escape single quotes in custom label descriptions
• Prevents malformed Python string literals in generated Enum
• Fixes labels with apostrophes like "Don't merge"

Diagram

flowchart LR
  A["Custom label description<br/>with single quote"] -->|escape single quotes| B["Properly formatted<br/>Python string literal"]
  B -->|generate Enum| C["Valid custom_labels_class"]

Loading

File Changes

1. pr_agent/algo/utils.py 🐞 Bug fix +1/-1

Escape single quotes in label descriptions

• Added .replace("'", "\\'") to escape single quotes in label descriptions
• Chained after existing newline replacement on line 960
• Prevents malformed Python string literals in generated Enum schema

pr_agent/algo/utils.py

---

[qodo-free-for-open-source-projects]

Code Review by Qodo

🐞 Bugs (2)   📘 Rule violations (1)   📎 Requirement gaps (0)

🐞\ ≡ Correctness (2) ⭐ New (2)

📘\ ⚙ Maintainability (1) ⭐ New (1)

1. v['description'] not validated ☑ 📘 ☼

Description

set_custom_labels() indexes v['description'] and calls string methods on it without validating
that v is a dict and description is a non-null string. A malformed or missing description in
user-provided custom_labels can raise KeyError/AttributeError and break label generation.

Code

pr_agent/algo/utils.py[959]

+        description = "'" + v['description'].strip('\n').replace('\n', '\\n').replace("'", "\\'") + "'"

Evidence

The checklist requires defensive validation of external/user-provided configuration values and
robust handling of edge cases. The changed line directly indexes v['description'] and calls
.strip()/.replace() with no type/null checks or safe access, so unexpected shapes in
custom_labels can crash at runtime.

Rule 3: Robust Error Handling
pr_agent/algo/utils.py[959-959]
Best Practice: Learned patterns

Agent prompt

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

## Issue description
`set_custom_labels()` assumes each `labels.items()` value is a dict with a string `description`, and calls string methods without null/type checks. This can crash on malformed user configuration.
## Issue Context
`labels` comes from `get_settings().get('custom_labels', {})`, which is user-configured input.
## Fix Focus Areas
- pr_agent/algo/utils.py[959-959]

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

---

2. New single-quoted literals added 📘 ⚙ ⭐

Description

The newly added string literals use single quotes (e.g., v.get('description', '') and
.strip('\n')) which conflicts with the repository convention in this checklist to prefer double
quotes. This can introduce style/lint inconsistencies in touched code.

Code

pr_agent/algo/utils.py[R959-960]

+        description = v.get('description', '') if isinstance(v, dict) else str(v)
+        description = "'" + description.strip('\n').replace('\n', '\\n').replace("'", "\\'") + "'"

Evidence

PR Compliance ID 8 requires adhering to the repository formatting conventions, including preferring
double quotes for string literals. The added code introduces multiple new single-quoted literals.

AGENTS.md
pr_agent/algo/utils.py[959-960]

Agent prompt

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

## Issue description
New code introduces single-quoted string literals where the checklist requires double quotes.

## Issue Context
To avoid style drift and potential formatter/linter noise, keep new literals aligned with the configured convention.

## Fix Focus Areas
- pr_agent/algo/utils.py[959-960]

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

---

3. Silent empty descriptions 🐞 ≡ ⭐

Description

set_custom_labels() now defaults to an empty string when a custom label entry has no
description, producing Label Enum members with empty values. This silently removes the label
meaning that the prompt instructs the LLM to rely on, making misconfigurations hard to detect and
reducing labeling quality.

Code

pr_agent/algo/utils.py[R959-960]

+        description = v.get('description', '') if isinstance(v, dict) else str(v)
+        description = "'" + description.strip('\n').replace('\n', '\\n').replace("'", "\\'") + "'"

Evidence

The PR changes behavior from v['description'] (fail-fast if missing) to v.get('description','')
(silent empty), and the prompts explicitly instruct the model to use the Enum value (description) to
understand label meaning; the repo’s custom label template also defines labels with a description
field.

pr_agent/algo/utils.py[941-966]
pr_agent/settings/pr_custom_labels.toml[17-34]
pr_agent/settings/custom_labels.toml[2-16]

Agent prompt

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

### Issue description
`set_custom_labels()` now silently uses an empty string when a label entry is missing `description`, producing Enum values that provide no meaning to the LLM. This should be treated as a configuration error (fail-fast) or at least fall back to a non-empty value (e.g., the label name) while logging a clear warning.

### Issue Context
The LLM prompt for labels explicitly says to use the `Label` value field to understand the label meaning; empty values degrade label selection and make misconfiguration hard to debug.

### Fix Focus Areas
- pr_agent/algo/utils.py[941-966]
- pr_agent/settings/pr_custom_labels.toml[17-34]

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

---

4. Invalid Enum member names 🐞 ≡ ⭐

Description

Enum member identifiers are derived from label names by only lowercasing and replacing spaces, so
labels containing characters like : generate malformed Label Enum definitions in the prompt
schema. This can break the schema shown to the LLM for documented label names such as `Main
topic:performance`.

Code

pr_agent/algo/utils.py[R962-963]

        variables["custom_labels_class"] += f"\n    {k.lower().replace(' ', '_')} = {description}"
        labels_minimal_to_labels_dict[k.lower().replace(' ', '_')] = k

Evidence

set_custom_labels() builds a Python-like Enum class definition string using `k.lower().replace('
', '_'), which does not sanitize characters like :` into valid identifiers. The repo documentation
explicitly gives Main topic:performance as a custom label example, which would produce an invalid
identifier (main_topic:performance) in the generated class injected into prompts.

pr_agent/algo/utils.py[954-966]
docs/docs/tools/describe.md[229-240]
pr_agent/settings/pr_custom_labels.toml[19-23]
pr_agent/settings/pr_description_prompts.toml[20-33]

Agent prompt

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

### Issue description
`set_custom_labels()` generates Enum member names from arbitrary label names but only replaces spaces, so labels containing `:`, `-`, `/`, leading digits, etc. can produce invalid identifiers and a malformed Enum schema in prompts.

### Issue Context
Docs show example label names with `:` (e.g., `Main topic:performance`). The generated `custom_labels_class` is embedded into both label-generation and description prompts, so malformed Enum text can confuse the model.

### Fix Focus Areas
- pr_agent/algo/utils.py[954-966]
- docs/docs/tools/describe.md[229-240]

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

---

ⓘ The new review experience is currently in Beta. Learn more

Previous review results

Review updated until commit 9314319

Results up to commit N/A

🐞 Bugs (0) 📘 Rule violations (0) 📎 Requirement gaps (0)

1. v['description'] not validated ☑ 📘 ☼

Description

set_custom_labels() indexes v['description'] and calls string methods on it without validating
that v is a dict and description is a non-null string. A malformed or missing description in
user-provided custom_labels can raise KeyError/AttributeError and break label generation.

Code

pr_agent/algo/utils.py[959]

+        description = "'" + v['description'].strip('\n').replace('\n', '\\n').replace("'", "\\'") + "'"

Evidence

The checklist requires defensive validation of external/user-provided configuration values and
robust handling of edge cases. The changed line directly indexes v['description'] and calls
.strip()/.replace() with no type/null checks or safe access, so unexpected shapes in
custom_labels can crash at runtime.

Rule 3: Robust Error Handling
pr_agent/algo/utils.py[959-959]
Best Practice: Learned patterns

Agent prompt

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

## Issue description
`set_custom_labels()` assumes each `labels.items()` value is a dict with a string `description`, and calls string methods without null/type checks. This can crash on malformed user configuration.
## Issue Context
`labels` comes from `get_settings().get('custom_labels', {})`, which is user-configured input.
## Fix Focus Areas
- pr_agent/algo/utils.py[959-959]

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

---

[qodo-free-for-open-source-projects]

Persistent review updated to latest commit 9314319