prism

skill
Security Audit
Warn
Health Warn
  • License — License: MIT
  • Description — Repository has a description
  • Active repo — Last push 0 days ago
  • Low visibility — Only 5 GitHub stars
Code Pass
  • Code scan — Scanned 12 files during light audit, no dangerous patterns found
Permissions Pass
  • Permissions — No dangerous permissions requested

No AI report is available for this listing yet.

SUMMARY

Session intelligence for Claude Code - find extra token usage, why your sessions fail, and fix them

README.md

Prism

Session intelligence for Claude Code. Find out why your sessions are burning tokens, and fix them.

PyPI version
License: MIT
Python 3.11+
Tests

PRISM demo

pip install prism-cc

The problem

Your Claude Code sessions are silently wasting tokens and you don't know why.

Running PRISM against real session data from a single machine found:

  • A project with 6738% CLAUDE.md re-read cost in one session — a 237-line file being re-read on every tool call
  • A project where CLAUDE.md re-reads consumed 480% of total session tokens — more tokens on instructions than on actual work
  • 4 migration file edits in a project that had a rule saying never to touch them — the rule existed, Claude ignored it
  • 5 consecutive tool failures in a single session with no diagnosis

None of this was visible before PRISM. The token counter just said you hit your limit.


What PRISM does

PRISM reads Claude Code's session files from ~/.claude/projects/ — the same files Claude Code writes automatically — and tells you three things:

  1. Why your tokens are disappearing — CLAUDE.md re-read costs, retry loops, compaction losses, sidechain waste
  2. Whether your CLAUDE.md rules are actually being followed — or silently ignored mid-session
  3. Exactly what to change — concrete diff recommendations, not generic advice

You keep using Claude Code exactly as normal. PRISM is the tool you run after.


Quick start

# Install
pip install prism-cc

# Analyze all your Claude Code projects
prism analyze

# Get specific CLAUDE.md fix recommendations
prism advise

# Open the full interactive dashboard
prism

# Watch a session live as it runs
prism watch

What you'll see

 Project                   Overall   Token Eff.  Tool Health  Ctx Hygiene  MD Adherence  Continuity
 myapp                      C+        D           B+           D            C             A
 ai-assistant               C         F           A            B            B+            A-
 data-pipeline              C+        C+          D            B            C+            B
 web-scraper                C+        D+          B            B+           B             A
 cli-tool                   B+        B+          A-           B+           A             A

Followed by the advisor:

╭──────────────────────────────────────────────────────────╮
│  PRISM ADVISOR — recommendations for myapp               │
╰──────────────────────────────────────────────────────────╯

  ✦ TRIM  (High impact — silent token drain every session)
    Remove lines 120–148: personality/tone instructions
    Claude Code's system prompt already handles this.
    These 29 lines cost tokens on every single tool call.

  ✦ RESTRUCTURE  (Reduce root-level re-read cost)
    Move 3 rules to subdirectory CLAUDE.md files:
    - "Use functional components only in React"
    - "Import from @/components, never relative paths"
    - "Run bun run typecheck after TypeScript changes"
    These only matter in src/ — loading them globally wastes
    tokens in every session that doesn't touch that directory.

Features

Dimension What PRISM measures
Token Efficiency CLAUDE.md re-read costs, cache hit patterns, compaction frequency
Tool Health Retry loops, edit-revert cycles, consecutive failures, interactive command hangs
Context Hygiene Compaction loss events, mid-task boundaries, sidechain fragmentation
CLAUDE.md Adherence Whether your rules are actually being followed — or ignored mid-session
Session Continuity Resume success rate, context loss on restart, truncated session files

How it works

You use Claude Code normally
         ↓
Claude Code writes session files to ~/.claude/projects/
         ↓
PRISM reads and analyzes those files
         ↓
Health scores  +  root cause diagnosis  +  CLAUDE.md diff

PRISM never touches Claude Code. It never modifies your sessions. It reads the JSONL files Claude Code already writes and surfaces what's inside them.


The CLAUDE.md re-read problem

Every tool call Claude Code makes re-reads your CLAUDE.md from the top of the context. A 200-line CLAUDE.md × 50 tool calls = 10,000 tokens spent on instructions, per session. If your CLAUDE.md has grown to include personality instructions, full documentation copies, or rules that only apply to one subdirectory — you're paying for all of it every time.

PRISM measures this exactly and tells you which lines are costing you the most.

The "Mr. Tinkleberry test" (from HN, 748 upvotes): put an absurd instruction in your CLAUDE.md. When Claude stops following it mid-session, your file has grown too long and adherence is degrading. PRISM automates this test across all your real sessions.


Commands

prism                          # Full interactive TUI dashboard
prism analyze                  # Rich-formatted health report, then exit
prism analyze --project ~/myproject   # One project only
prism analyze --json           # JSON output for scripting
prism advise                   # CLAUDE.md recommendations
prism advise --apply           # Write recommendations (with confirmation)
prism replay <session-id>      # Scrub through a session timeline
prism watch                    # Live dashboard for the running session
prism projects                 # List all projects with session counts

Install

# pip
pip install prism-cc

# pipx (recommended — isolated install)
pipx install prism-cc

# from source
git clone https://github.com/jakeefr/prism
cd prism
pip install -e .

Requires Python 3.11+. No Claude API key needed — reads local files only. Works on macOS, Linux, and Windows.


Contributing

Issues and PRs welcome. If you run prism analyze and find something interesting in your own session data, open an issue — real-world patterns help improve the detection logic.

git clone https://github.com/jakeefr/prism
cd prism
uv sync --dev
uv run pytest

License

MIT — do whatever you want with it.


PRISM doesn't send anything anywhere. All analysis runs locally against files already on your machine.

Reviews (0)

No results found