On this page
Model Selection and Triggers
This page explains how to choose the right LLM model for each analysis role and how to use CEL expressions to control when Pipelines-as-Code triggers analysis. Use model selection to balance cost and capability, and use CEL triggers to limit analysis to the events that matter.
Model Selection
Each analysis role can specify a different model. Choosing the right model lets you balance cost against analysis depth. If you do not specify a model, Pipelines-as-Code uses provider-specific defaults:
- OpenAI:
gpt-5.4-mini - Gemini:
gemini-3.1-flash-lite-preview
Specifying Models
You can use any model name that your chosen LLM provider supports. Consult the provider’s documentation for available models:
- OpenAI Models: https://platform.openai.com/docs/models
- Gemini Models: https://ai.google.dev/gemini-api/docs/models/gemini
Example: Assigning Different Models per Role
The following example shows how to assign a different model to each analysis role, matching the model’s capability to the task’s complexity:
settings:
ai:
enabled: true
provider: "openai"
secret_ref:
name: "openai-api-key"
key: "token"
roles:
# Use the most capable model for complex analysis
- name: "security-analysis"
model: "gpt-5"
prompt: "Analyze security failures..."
# Use default model (gpt-5.4-mini) for general analysis
- name: "general-failure"
# No model specified - uses provider default
prompt: "Analyze this failure..."
# Use the most economical model for quick checks
- name: "quick-check"
model: "gpt-5-nano"
prompt: "Quick diagnosis..."CEL Expressions for Triggers
By default, Pipelines-as-Code runs LLM analysis for completed PipelineRuns (both successful and failed). CEL (Common Expression Language) expressions in the on_cel field let you refine this behavior – for example, restricting analysis to a specific branch, only successful runs, or only failed runs.
If you omit on_cel, the role executes for all completed PipelineRuns.
Each role is checked on its own, so one PipelineRun can trigger several roles, each with its own model:
flowchart LR
%% title: How each analysis role decides whether to run
Run@{ img: "/icons/pipeline.svg", label: "Finished run", w: 48, h: 48, constraint: "on" }
Has@{ img: "/icons/decision.svg", label: "on_cel set?", w: 48, h: 48, constraint: "on" }
Done[Completed?]
Cel@{ img: "/icons/filter.svg", label: "Evaluate CEL", w: 48, h: 48, constraint: "on" }
Model@{ img: "/icons/model.svg", label: "Role model", w: 48, h: 48, constraint: "on" }
Skip@{ img: "/icons/stop.svg", label: "Skipped", w: 48, h: 48, constraint: "on" }
Err@{ img: "/icons/warning.svg", label: "Role failed", w: 48, h: 48, constraint: "on" }
Run --> Has
Has unset@-->|no| Done
Has set@-->|yes| Cel
Done done@-->|yes| Model
Cel true@-->|true| Model
Cel false@-->|false| Skip
Cel bad@-->|error| Err
%% play: No on_cel
%% skip Cel Skip Err
%% Run -> Has : role
%% Has -> Done @unset | Without on_cel, the role runs when the PipelineRun succeeded or failed
%% Done -> Model @done : gpt-5.4-mini | The role uses its model, or the provider default when none is set
%% play: Expression matches
%% skip Done Skip Err
%% Run -> Has : role
%% Has -> Cel @set : body | The expression sees body.pipelineRun, body.repository and body.event
%% Cel -> Model @true : gpt-5
%% play: Expression is false
%% skip Done Model Err
%% Run -> Has : role
%% Has -> Cel @set
%% Cel -> Skip @false | The role is skipped quietly
%% play: Broken expression
%% skip Done Model Skip
%% Run -> Has : role
%% Has -> Cel @set
%% Cel -> Err @bad | An invalid expression, or one that does not return true or false, fails that role onlyOverriding the Default Behavior
To run analysis for all completed PipelineRuns (both successful and failed), set on_cel: 'true':
roles:
- name: "pipeline-summary"
prompt: "Generate a summary of this pipeline run..."
on_cel: 'true' # Runs for all completed pipeline runs
output: "pr-comment"This is useful when you want to:
- Generate summaries for every PipelineRun
- Track metrics for successful runs
- Post automated messages on successful builds
- Report on build performance
Example CEL Expressions
# Run on ALL pipeline runs (overrides default completed-only behavior)
on_cel: 'true'
# Only on successful runs (e.g., for generating success reports)
on_cel: 'body.pipelineRun.status.conditions[0].reason == "Succeeded"'
# Only on pull requests
on_cel: 'body.event.event_type == "pull_request"'
# Only on main branch
on_cel: 'body.event.base_branch == "main"'
# Only on default branch (works across repos with different default branches)
on_cel: 'body.event.base_branch == body.event.default_branch'
# Skip analysis for bot users
on_cel: 'body.event.sender != "dependabot[bot]"'
# Only for PRs with specific labels
on_cel: '"needs-review" in body.event.pull_request_labels'
# Only when triggered by comment
on_cel: 'body.event.trigger_comment.startsWith("/analyze")'
# Combine conditions
on_cel: 'body.pipelineRun.status.conditions[0].reason == "Failed" && body.event.event_type == "pull_request"'Available CEL Context Fields
The following tables list all fields you can reference in on_cel expressions. Pipelines-as-Code populates these fields from the PipelineRun status, the Repository CR, and the Git provider event.
Top-Level Context
| Field | Type | Description |
|---|---|---|
body.pipelineRun | object | Full PipelineRun object with status and metadata |
body.repository | object | Full Repository CR object |
body.event | object | Event information (see Event Fields below) |
pac | map[string]string | PAC parameters map |
Event Fields (body.event.*)
Event Type and Trigger:
| Field | Type | Description | Example |
|---|---|---|---|
event_type | string | Event type from provider | "pull_request", "push", "Merge Request Hook" |
trigger_target | string | Normalized trigger type across providers | "pull_request", "push" |
Branch and Commit Information:
| Field | Type | Description | Example |
|---|---|---|---|
sha | string | Commit SHA | "abc123def456..." |
sha_title | string | Commit title/message | "feat: add new feature" |
base_branch | string | Target branch for PR (or branch for push) | "main" |
head_branch | string | Source branch for PR (or branch for push) | "feature-branch" |
default_branch | string | Default branch of the repository | "main" or "master" |
Repository Information:
| Field | Type | Description | Example |
|---|---|---|---|
organization | string | Organization/owner name | "my-org" |
repository | string | Repository name | "my-repo" |
URLs:
| Field | Type | Description | Example |
|---|---|---|---|
url | string | Web URL to repository | "https://github.com/org/repo" |
sha_url | string | Web URL to commit | "https://github.com/org/repo/commit/abc123" |
base_url | string | Web URL to base branch | "https://github.com/org/repo/tree/main" |
head_url | string | Web URL to head branch | "https://github.com/org/repo/tree/feature" |
User Information:
| Field | Type | Description | Example |
|---|---|---|---|
sender | string | User who triggered the event | "user123", "dependabot[bot]" |
Pull Request Fields (only populated for PR events):
| Field | Type | Description | Example |
|---|---|---|---|
pull_request_number | int | PR/MR number | 42 |
pull_request_title | string | PR/MR title | "Add new feature" |
pull_request_labels | []string | List of PR/MR labels | ["enhancement", "needs-review"] |
Comment Trigger Fields (only when triggered by comment):
| Field | Type | Description | Example |
|---|---|---|---|
trigger_comment | string | Comment that triggered the run | "/test", "/retest" |
Webhook Fields:
| Field | Type | Description | Example |
|---|---|---|---|
target_pipelinerun | string | Target PipelineRun for incoming webhooks | "my-pipeline-run" |
Excluded Fields
Pipelines-as-Code intentionally excludes the following fields from the CEL context for security and architectural reasons:
event.Provider– Contains sensitive API tokens and webhook secrets.event.Request– Contains raw HTTP headers and payload, which may include secrets.event.InstallationID,AccountID,GHEURL,CloneURL– Provider-specific internal identifiers and URLs.event.SourceProjectID,TargetProjectID– GitLab-specific internal identifiers.event.State– Internal state management fields.event.Event– Raw provider event object (already represented in the structured fields above).
Output Destinations
Output destinations control where Pipelines-as-Code posts the LLM analysis results.
PR Comment
Posts analysis as a comment on the pull request:
output: "pr-comment"Benefits of PR comments:
- Visible to all developers working on the pull request.
- Pipelines-as-Code can update the comment with new analysis on subsequent runs.
- Easy to discuss and follow up on directly in the PR conversation.
check-run (GitHub check runs) and annotation (PipelineRun annotations) are planned for future releases.