ci: update config to be less verbose

[maurerle]

User description

Trying to make the pr agent less verbose from the feedback of the dev meeting.

---

PR Type

Enhancement

---

Description

• Upgrade PR Agent action version

• Modernize PR Agent environment configuration

• Enable labels and description comments

• Disable auto-review and final updates

---

Diagram Walkthrough

flowchart LR
  A["PR Agent workflow"] 
  B["Action v0.39.0"]
  C["Updated config keys"]
  D["Labels and description comments"]
  E["Auto-review disabled"]
  A -- "uses" --> B
  A -- "configures" --> C
  C -- "publishes" --> D
  C -- "turns off" --> E

Loading

File Walkthrough

	Relevant files
Configuration changes	pr-agent.yml Update PR Agent workflow settings                                                .github/workflows/pr-agent.yml Upgrades the-pr-agent/pr-agent from v0.36.1 to v0.39.0. Replaces older CONFIG__... environment keys with config... settings. Enables label publishing, description comments, and relevant configuration output. Disables automatic review and final update messaging. +9/-4

---

---

🛠️ Relevant configurations:

These are the relevant configurations for this tool:

[config]

custom_model_max_tokens: 128000
publish_description_as_comment: True
output_relevant_configurations: True
model: openai/gpt-5.5
publish_labels: True
fallback_models: ['openai/gpt-5.4', 'openai/openai-gpt-oss-120b']
final_update_message: False
git_provider: github
publish_output: True
publish_output_progress: True
progress_gif_url: 
progress_gif_width: 48
verbosity_level: 0
use_extra_bad_extensions: False
log_level: DEBUG
use_wiki_settings_file: True
use_repo_settings_file: True
use_global_settings_file: True
extra_config_url: 
disable_auto_feedback: False
ai_timeout: 120
custom_reasoning_model: False
response_language: en-US
repo_context_files: ['AGENTS.md']
repo_context_from_default_branch: True
repo_context_max_lines: 500
max_description_tokens: 500
max_commits_tokens: 500
max_model_tokens: 32000
model_token_count_estimate_factor: 0.3
patch_extension_skip_types: ['.md', '.txt']
allow_dynamic_context: True
max_extra_lines_before_dynamic_context: 10
patch_extra_lines_before: 5
patch_extra_lines_after: 1
cli_mode: False
large_patch_policy: clip
duplicate_prompt_examples: False
seed: -1
temperature: 0.2
ignore_pr_title: ['^\\[Auto\\]', '^Auto']
ignore_pr_target_branches: []
ignore_pr_source_branches: []
ignore_pr_labels: []
ignore_pr_authors: []
ignore_repositories: []
ignore_language_framework: []
restricted_mode: False
is_auto_command: True
enable_ai_metadata: False
reasoning_effort: medium
enable_claude_extended_thinking: False
extended_thinking_budget_tokens: 2048
extended_thinking_max_output_tokens: 4096
claude_extended_thinking_models_override: []
extract_issue_from_branch: True
branch_issue_regex: 
enable_custom_labels: False

[pr_description]

publish_labels: False
add_original_user_description: True
generate_ai_title: False
use_bullet_points: True
extra_instructions: 
enable_pr_type: True
final_update_message: False
enable_help_text: False
enable_help_comment: False
enable_pr_diagram: True
publish_description_as_comment: False
publish_description_as_comment_persistent: True
enable_semantic_files_types: True
collapsible_file_list: adaptive
collapsible_file_list_threshold: 6
inline_file_summary: False
use_description_markers: False
enable_large_pr_handling: True
include_generated_by_header: True
max_ai_calls: 4
async_ai_calls: True

[github-actions]

PR Code Suggestions ✨

Latest suggestions up to b0b2a42

Explore these optional code suggestions:

Category	Suggestion	Impact
Possible issue	Quote environment boolean Quote the boolean value in env so GitHub Actions and the action receive it consistently as a string. Leaving false unquoted can be parsed as a YAML boolean, which may break workflow validation or config parsing depending on the runner/action handling. .github/workflows/pr-agent.yml [31] -config.final_update_message: false +config.final_update_message: "false" Suggestion importance[1-10]: 8 __ Why: The suggestion is accurate because env values in GitHub Actions should be strings, and unquoted false may be parsed as a YAML boolean. Quoting config.final_update_message avoids possible workflow validation or action configuration issues.	Medium

---

🛠️ Relevant configurations:

These are the relevant configurations for this tool:

[config]

enable_ai_metadata: False
custom_model_max_tokens: 128000
publish_description_as_comment: True
output_relevant_configurations: True
model: openai/gpt-5.5
publish_labels: True
fallback_models: ['openai/gpt-5.4', 'openai/openai-gpt-oss-120b']
final_update_message: False
git_provider: github
publish_output: True
publish_output_progress: True
progress_gif_url: 
progress_gif_width: 48
verbosity_level: 0
use_extra_bad_extensions: False
log_level: DEBUG
use_wiki_settings_file: True
use_repo_settings_file: True
use_global_settings_file: True
extra_config_url: 
disable_auto_feedback: False
ai_timeout: 120
custom_reasoning_model: False
response_language: en-US
repo_context_files: ['AGENTS.md']
repo_context_from_default_branch: True
repo_context_max_lines: 500
max_description_tokens: 500
max_commits_tokens: 500
max_model_tokens: 32000
model_token_count_estimate_factor: 0.3
patch_extension_skip_types: ['.md', '.txt']
allow_dynamic_context: True
max_extra_lines_before_dynamic_context: 10
patch_extra_lines_before: 5
patch_extra_lines_after: 1
cli_mode: False
large_patch_policy: clip
duplicate_prompt_examples: False
seed: -1
temperature: 0.2
ignore_pr_title: ['^\\[Auto\\]', '^Auto']
ignore_pr_target_branches: []
ignore_pr_source_branches: []
ignore_pr_labels: []
ignore_pr_authors: []
ignore_repositories: []
ignore_language_framework: []
restricted_mode: False
is_auto_command: True
reasoning_effort: medium
enable_claude_extended_thinking: False
extended_thinking_budget_tokens: 2048
extended_thinking_max_output_tokens: 4096
claude_extended_thinking_models_override: []
extract_issue_from_branch: True
branch_issue_regex: 
enable_custom_labels: False

[pr_code_suggestions]

commitable_code_suggestions: False
dual_publishing_score_threshold: -1
focus_only_on_problems: True
extra_instructions: 
enable_help_text: False
enable_chat_text: False
persistent_comment: True
max_history_len: 4
publish_output_no_suggestions: True
suggestions_score_threshold: 0
new_score_mechanism: True
new_score_mechanism_th_high: 9
new_score_mechanism_th_medium: 7
auto_extended_mode: True
num_code_suggestions_per_chunk: 3
max_number_of_calls: 3
parallel_calls: True
final_clip_factor: 0.8
decouple_hunks: False
demand_code_suggestions_self_review: False
code_suggestions_self_review_text: **Author self-review**: I have reviewed the PR code suggestions, and addressed the relevant ones.
approve_pr_on_self_review: False
fold_suggestions_on_self_review: True

---

Previous suggestions

Suggestions up to commit 446078e

Category	Suggestion	Impact
Possible issue	Quote boolean environment value Quote config.final_update_message so it is passed as an environment variable string instead of a YAML boolean. GitHub Actions env values are safest as strings, and an unquoted boolean can cause workflow validation or action config parsing issues. .github/workflows/pr-agent.yml [31] -config.final_update_message: false +config.final_update_message: "false" Suggestion importance[1-10]: 8 __ Why: This is a valid correctness issue because unquoted false under GitHub Actions env may be parsed as a YAML boolean instead of an environment variable string. Quoting config.final_update_message is a small but important fix that can prevent workflow validation or action configuration problems.	Medium

Suggestions up to commit ee2cfbf

Category	Suggestion	Impact
Possible issue	Quote environment boolean values Quote config.final_update_message so GitHub Actions treats it as an environment variable string. Leaving it as a YAML boolean can make the workflow invalid or pass an unexpected value to the action. .github/workflows/pr-agent.yml [30] -config.final_update_message: false +config.final_update_message: "false" Suggestion importance[1-10]: 8 __ Why: This is a valid workflow correctness issue because values under env should be strings, and unquoted false may be interpreted as a YAML boolean. The suggested change accurately fixes config.final_update_message without altering the intended value.	Medium
General	Disable configuration output Disable config.output_relevant_configurations if the goal is to reduce PR-Agent verbosity. Keeping it enabled can add extra configuration output to runs or comments and may expose unnecessary setup details. .github/workflows/pr-agent.yml [29] -config.output_relevant_configurations: "true" +config.output_relevant_configurations: "false" Suggestion importance[1-10]: 2 __ Why: The suggestion is technically applicable to config.output_relevant_configurations, but it is speculative and may contradict the PR's apparent intent to enable this PR-Agent output. It is a low-impact preference unless reducing verbosity or hiding configuration details is an explicit goal.	Low

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 82.13%. Comparing base (09c819e) to head (b0b2a42).
⚠️ Report is 12 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #834   +/-   ##
=======================================
  Coverage   82.13%   82.13%           
=======================================
  Files          56       56           
  Lines        9072     9072           
=======================================
  Hits         7451     7451           
  Misses       1621     1621           

Flag	Coverage Δ
pytest	82.13% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[reinecfi]

A little less verbosity sound reasonable.

[maurerle]

somehow the label creation is lost now - hm.
For now the PR-Agent does not allow all choices what to have and what not :(

config.output_relevant_configurations shows the existing config, but not which model was used.
i created two issues for pr-agent:

• add information about which model is actually used The-PR-Agent/pr-agent#2514
• disable AI description - show only filetree and diagram The-PR-Agent/pr-agent#2516

[maurerle]

For now, I think that it is rather fine to have both blocks (code suggestions and reviewer guide).
In here: #839 (comment) it found important fixes.

So I would just like to fix the two open issues (not overwriting the issue desc that much and providing more information about used models/tokens)