• There are 352 files with 113,446 lines of code.
• 31 very long files (54,218 lines of code)
• 37 long files (25,294 lines of code)
• 62 medium size files (19,138 lines of code)
• 58 small files (8,403 lines of code)
• 164 very small files (6,393 lines of code)

File	# lines	≈ tokens to read	# changes (1y)	≈ lines edited per change	lines read per line changed
run_state.py in src/agents	4676	~30k–70k	63	~100	~30
run.py in src/agents	2405	~15k–50k	86	~50	~50
run_loop.py in src/agents/run_internal	2632	~15k–50k	76	~70	~50
turn_resolution.py in src/agents/run_internal	3602	~20k–70k	44	~100	~30
tool_execution.py in src/agents/run_internal	2414	~15k–50k	47	~70	~30
openai_responses.py in src/agents/models	2002	~10k–30k	56	~30	~70
tool.py in src/agents	1964	~10k–30k	54	~50	~50
server.py in src/agents/mcp	1947	~10k–30k	54	~70	~30
openai_realtime.py in src/agents/realtime	1870	~10k–30k	51	~50	~30
session.py in src/agents/realtime	1841	~10k–30k	46	~50	~30
session_persistence.py in src/agents/run_internal	1541	~10k–30k	32	~70	~20
docker.py in src/agents/sandbox/sandboxes	1881	~10k–30k	22	~100	~15
unix_local.py in src/agents/sandbox/sandboxes	1397	~7k–20k	29	~70	~20
any_llm_model.py in src/agents/extensions/models	1374	~7k–20k	29	~70	~20
agent.py in src/agents	832	~5k–15k	47	~20	~30
result.py in src/agents	883	~5k–15k	38	~30	~30
advanced_sqlite_session.py in src/agents/extensions/memory	1147	~7k–20k	27	~100	~10
chatcmpl_stream_handler.py in src/agents/models	1093	~7k–20k	28	~50	~20
litellm_model.py in src/agents/extensions/models	794	~5k–15k	37	~20	~30
sandbox.py in src/agents/extensions/sandbox/modal	1564	~10k–30k	18	~150	~10

Tokens are lines × 7 to 14 (analysis.fileReadsForChanges). Lines edited per change are the file's lines added and deleted divided by its commits over the whole history. Agents read about 2,000 lines at a time: when a 17,000-line file was split into modules, the tokens an agent read for the same change fell by 83% (The Economic Benefit of Refactoring, 2026).

There are 12 files with lines longer than 120 characters. In total, there are 14 long lines.

File	# lines	# units	# long lines
run.ps1 in .agents/skills/code-change-verification/scripts	172	-	3
openai.yaml in .agents/skills/code-change-verification/agents	4	-	1
openai.yaml in .agents/skills/final-release-review/agents	4	-	1
openai.yaml in .agents/skills/test-coverage-improver/agents	4	-	1
openai.yaml in .agents/skills/implementation-kickoff/agents	6	-	1
openai.yaml in .agents/skills/implementation-final-review/agents	6	-	1
openai.yaml in .agents/skills/sensitive-logging-audit/agents	4	-	1
openai.yaml in .agents/skills/implementation-strategy/agents	4	-	1
openai.yaml in .agents/skills/pr-draft-summary/agents	4	-	1
openai.yaml in .agents/skills/maintainer-review/agents	4	-	1
openai.yaml in .agents/skills/runtime-behavior-probe/agents	6	-	1
openai.yaml in .agents/skills/examples-run-analysis/agents	4	-	1

• File size measurements show the distribution of size of files.
• Files are classified in four categories based on their size (lines of code): 1-100 (very small files), 101-200 (small files), 201-500 (medium size files), 501-1000 (long files), 1001+(very long files).
• It is a good practice to keep files small. Long files may become "bloaters", code that have increased to such gargantuan proportions that they are hard to work with.