Imported from oimiragieo/agent-studio (.claude/skills/context-attribution/SKILL.md). Install upstream with npx skills add oimiragieo/agent-studio --skill context-attribution. Copyright stays with the author.

Context Attribution

Estimate per-turn token attribution across 6 categories in Claude Code sessions. Based on claude-devtools visible context tracker.

When to Invoke

Skill({ skill: 'context-attribution' });

Use when: context pressure is high, optimizing CLAUDE.md sizes, understanding which tool calls consume the most tokens, debugging context overflow.

The 6 Categories

Workflow

Step 1: Load Session JSONL

# Find most recent session
SESSION=$(ls -t ~/.claude/projects/$(pwd | sed 's|/|-|g; s|^-||')/*.jsonl | head -1)

Step 2: Extract Per-Turn Token Data

For each message, classify into one of the 6 categories and estimate tokens:

# Count user messages (Category 6)
grep '"role":"user"' "$SESSION" | grep -v '"tool_result"' | wc -l

# Count tool results (Categories 2-3)
grep '"type":"tool_result"' "$SESSION" | wc -l

# Count assistant output (Category 4)
grep '"role":"assistant"' "$SESSION" | wc -l

# Check for team messages (Category 5)
grep 'teammate-message' "$SESSION" | wc -l

Step 3: Estimate Tokens Per Category

Use the usage field from each assistant turn for accurate counts. Fall back to chars/4 when unavailable.

Step 4: Output Attribution Table

Turn | CLAUDE.md | Files | Tools | AI Out | Team | User | Total
-----|-----------|-------|-------|--------|------|------|------
  1  |   12,400  |     0 |     0 |    800 |    0 |  200 | 13,400
  2  |        0  | 3,200 | 1,500 |  2,100 |    0 |  150 |  6,950
...  |      ...  |   ... |   ... |    ... |  ... |  ... |   ...

Step 5: Identify Top Consumers

Report which category consumes the most tokens and suggest optimizations (e.g., reduce CLAUDE.md size, compress tool outputs).

Memory Protocol (MANDATORY)

Before starting: ```bash cat .claude/context/memory/learnings.md cat .claude/context/memory/decisions.md ```

After completing:

• New pattern -> `.claude/context/memory/learnings.md`
• Issue found -> `.claude/context/memory/issues.md`
• Decision made -> `.claude/context/memory/decisions.md`

ASSUME INTERRUPTION: Your context may reset. If it's not in memory, it didn't happen.