Raise diff size defaults for 1M context models

The previous DEFAULT_MAX_DIFF_CHARS (400k chars / ~100k tokens) was sized
for 200k-token context models. Both Opus 4.7 and Sonnet 4.6 now expose a
1M context window, so double the default to 800k chars (~200k tokens),
which still leaves ~5x context headroom for system prompt, tools,
thinking, and output.

Also wire get_pr_data's fallback default to DEFAULT_MAX_DIFF_CHARS so
the constant is the single source of truth, refresh the action.yml and
README guidance for 1M-context defaults, and clarify the
PROMPT_TOKEN_LIMIT comment (it caps filter/validator output, not input).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Author: matej

README.md

@@ -111,9 +111,9 @@ This action is not hardened against prompt injection attacks and should only be
 | `comment-pr` | Whether to comment on PRs with findings | `true` | No |
 | `upload-results` | Whether to upload results as artifacts | `true` | No |
 | `exclude-directories` | Comma-separated list of directories to exclude from scanning | None | No |
-| `claude-model` | Claude [model name](https://docs.anthropic.com/en/docs/about-claude/models/overview#model-names) to use. Defaults to Opus 4.7. For large PRs (>400k char diffs), consider using `claude-sonnet-4-6` (1M context). | `claude-opus-4-7` | No |
+| `claude-model` | Claude [model name](https://docs.anthropic.com/en/docs/about-claude/models/overview#model-names) to use. Defaults to Opus 4.7 (1M context). For very large PRs or to reduce cost, consider `claude-sonnet-4-6` (also 1M context, faster and cheaper). | `claude-opus-4-7` | No |
 | `claudecode-timeout` | Timeout for ClaudeCode analysis in minutes | `20` | No |
-| `max-diff-chars` | Maximum diff characters to include in prompt. Set to `0` for agentic mode (Claude uses git commands to explore). See [Diff Size Configuration](#diff-size-configuration) below. | `400000` | No |
+| `max-diff-chars` | Maximum diff characters to include in prompt. Set to `0` for agentic mode (Claude uses git commands to explore). See [Diff Size Configuration](#diff-size-configuration) below. | `800000` | No |
 | `max-diff-lines` | **[DEPRECATED]** Use `max-diff-chars` instead. Converts lines to chars (line × 80). | None | No |
 | `run-every-commit` | Run ClaudeCode on every commit (skips cache check). Warning: May increase false positives on PRs with many commits. **Deprecated**: Use `trigger-on-commit` instead. | `false` | No |
 | `trigger-on-open` | Run review when PR is first opened | `true` | No |
@@ -146,7 +146,7 @@ The action handles PRs of any size using three review modes:
 1. **Full Diff Mode** (default for small PRs)
    - Entire diff embedded in prompt
    - Fastest and most comprehensive
-   - Works for diffs up to ~400k characters
+   - Works for diffs up to ~800k characters
 
 2. **Partial Diff Mode** (automatic for large PRs)
    - First N files embedded in prompt
@@ -160,21 +160,20 @@ The action handles PRs of any size using three review modes:
 
 #### Configuration
 
-**`max-diff-chars`** - Maximum diff characters to embed (default: 400,000)
+**`max-diff-chars`** - Maximum diff characters to embed (default: 800,000)
 
 ```yaml
-# Default: 400k chars (fits in 200k token models)
+# Default: 800k chars (~200k tokens, fits comfortably in 1M context models)
 - uses: PSPDFKit-labs/nutrient-code-review@main
   with:
     claude-api-key: ${{ secrets.CLAUDE_API_KEY }}
-    max-diff-chars: 400000  # ~100k tokens
+    max-diff-chars: 800000
 
-# Large PRs: Use 1M context model with higher limit
+# Very large PRs: push the embed budget higher (still within 1M context)
 - uses: PSPDFKit-labs/nutrient-code-review@main
   with:
     claude-api-key: ${{ secrets.CLAUDE_API_KEY }}
-    claude-model: claude-sonnet-4-6  # 1M context
-    max-diff-chars: 800000  # ~200k tokens
+    max-diff-chars: 1600000  # ~400k tokens
 
 # Always use agentic mode (no embedded diff)
 - uses: PSPDFKit-labs/nutrient-code-review@main
@@ -187,9 +186,9 @@ The action handles PRs of any size using three review modes:
 
 | Diff Size | Recommended Model | Context Window |
 |-----------|-------------------|----------------|
-| < 400k chars | `claude-opus-4-7` (default) | 1M tokens |
-| 400k - 800k chars | `claude-sonnet-4-6` | 1M tokens |
-| > 800k chars | Set `max-diff-chars: 0` (agentic mode) | Any model |
+| < 800k chars | `claude-opus-4-7` (default) | 1M tokens |
+| 800k - 1.6M chars | `claude-opus-4-7` or `claude-sonnet-4-6` with raised `max-diff-chars` | 1M tokens |
+| > 1.6M chars | Set `max-diff-chars: 0` (agentic mode) | Any model |
 
 **Backward Compatibility:**
 

action.yml

@@ -71,16 +71,15 @@ inputs:
 
   max-diff-chars:
     description: |
-      Maximum diff characters to embed in prompt (default: 400000 = 400k chars).
-      Larger diffs use agentic file reading instead. Set to 0 to always use agentic mode.
+      Maximum diff characters to embed in prompt (default: 800000 = 800k chars,
+      ~200k tokens). Larger diffs use agentic file reading instead. Set to 0 to
+      always use agentic mode.
 
-      IMPORTANT: For large limits (>400k), use a model with larger context like:
-      - claude-sonnet-4-6 (1M context) for diffs up to 800k chars
-      - Set via 'claude-model' input parameter
-
-      Note: ~400k chars fits comfortably in the default Opus 4.7 model (1M context).
+      Note: 800k chars fits comfortably in the default Opus 4.7 model (1M
+      context). To embed even larger diffs, raise this value — both Opus 4.7
+      and Sonnet 4.6 have 1M context.
     required: false
-    default: '400000'
+    default: '800000'
 
   max-diff-lines:
     description: |

claudecode/constants.py

@@ -11,10 +11,10 @@
 RATE_LIMIT_BACKOFF_MAX = 30  # Maximum backoff time for rate limits
 
 # Token Limits
-PROMPT_TOKEN_LIMIT = 16384  # 16k tokens max for claude-opus-4
+PROMPT_TOKEN_LIMIT = 16384  # Output cap for filter/validator API calls
 
 # Diff Construction Limits
-DEFAULT_MAX_DIFF_CHARS = 400000  # 400k characters (suitable for 200k token models)
+DEFAULT_MAX_DIFF_CHARS = 800000  # 800k characters (~200k tokens; fits comfortably in 1M context models)
 # Conversion factor for deprecated MAX_DIFF_LINES -> MAX_DIFF_CHARS
 CHARS_PER_LINE_ESTIMATE = 80  # Average characters per line for conversion
 

claudecode/github_action_audit.py

@@ -128,7 +128,7 @@ def __init__(self):
             print(f"[Debug] User excluded directories: {user_excluded_dirs}", file=sys.stderr)
         print(f"[Debug] Total excluded directories: {self.excluded_dirs}", file=sys.stderr)
 
-    def get_pr_data(self, repo_name: str, pr_number: int, max_diff_chars: int = 400000) -> Dict[str, Any]:
+    def get_pr_data(self, repo_name: str, pr_number: int, max_diff_chars: int = DEFAULT_MAX_DIFF_CHARS) -> Dict[str, Any]:
         """Get PR metadata and construct diff in one pass with early termination.
 
         Fetches files page-by-page while building the diff. Stops fetching when