Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add instructions/jaredgizersky/stash/claude-mdgit clone --depth 1 https://github.com/jaredgizersky/stashWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/jaredgizersky/stash/claude-md)<a href="https://agentmods.dev/instructions/jaredgizersky/stash/claude-md"><img src="https://agentmods.dev/badge/instructions/jaredgizersky/stash/claude-md.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.01201 | $0.01201 |
| Opus 5 | $0.00600 | $0.00600 |
| Sonnet 5 | $0.00240 | $0.00240 |
| Haiku 4.5 | $0.00120 | $0.00120 |
Grade A, and why
stash CLAUDE.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 83 lines — stays where its author put it; the contents beside it link to each section on GitHub.
stash
CLI/TUI for managing Claude Code and Codex sessions. Lets you name, bookmark ("stash"), browse, preview transcripts, and resume sessions.
Architecture
cmd/stash/main.go CLI entry point (cobra). Subcommands: (default TUI), hook, list
internal/claude/sessions.go Session loading from ~/.claude/projects/*/*.jsonl. Parallel scanning with file-level cache (~/.stash/session-cache.json)
internal/claude/cache.go mtime+size-based cache for scanned JSONL metadata. Bump cacheVersion const if Session struct changes
internal/claude/active.go Reads ~/.claude/sessions/*.json for live processes (PID, alive check, cwd)
internal/codex/sessions.go Codex session loading from ~/.codex/state_5.sqlite (threads table). Pure-Go SQLite via modernc.org/sqlite
internal/codex/transcript.go Codex JSONL transcript reader (response_item format → TranscriptEntry)
internal/store/store.go Stash index (~/.stash/index.json) — the ledger of explicitly stashed sessions
internal/tui/tui.go Bubble Tea TUI. Three tabs (Stash/History/Active), unified update handler, transcript preview
How Claude sessions work
- Sessions stored as JSONL at
~/.claude/projects/<encoded-cwd>/<session-id>.jsonl - Encoding: absolute path with every non-alnum char replaced by
-(lossy — cannot be reversed) - Some project dirs have
sessions-index.jsonbut it's incomplete (only covers ~5% of sessions). We ignore it and scan all JONLs directly. - Session names set via
/renameorclaude -nappear as{"type": "agent-name", "agentName": "..."}entries in the JSONL - Active sessions:
~/.claude/sessions/<pid>.json— contains pid, sessionId, cwd, name. File keyed by PID. - Resume:
claude --resume <session-id>in the correct cwd
Data flow
LoadAllSessions()scans all JSONL files across all project dirs, parallelized, cached by mtime+sizeLoadActiveSessions()reads~/.claude/sessions/*.json, checks PIDs alive viakill -0LinkTranscripts()matches active sessions to their transcript Session objects- Stash index (
~/.stash/index.json) marks which sessions are "stashed" — populated by the hook - TUI enriches sessions with stash names, renders three tabs
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 5d ago First seen · 83 lines · 1,201 tokens per session scan A 9a6e34df77b2
stash CLAUDE.md is an instructions file published in the GitHub repository jaredgizersky/stash (11 stars, last pushed 4mo ago), licensed MIT. It adds 1,201 tokens to every session, about $0.0060 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
terminal AGENTS.md
Instructions for browser-use/terminal, covering agent notes, rust rewrite verification loop and terminal ui testing standard.
hex1b AGENTS.md
Instructions for mitchdenny/hex1b, covering ai coding agent guidelines for hex1b, 🛠️ available skills, 📋 project overview, key technologies and repository layout.
waveloom AGENTS.md
AGENTS.md instructions for Menfre01/waveloom, covering waveloom, 项目概要, 编码规范, 代码审查 and 开发流程.
unifly AGENTS.md
Instructions for hyperb1iss/unifly, covering agents.md, what this repository is, quick command reference, the canonical "before commit" gate and individual gates.
lazyazure AGENTS.md
AGENTS.md instructions for matsest/lazyazure, covering lazyazure - agent guidelines, project overview, critical guidelines, 1. tui development with gocui and 2. mutex best practices.
r3bl-open-core AGENTS.md
AGENTS.md instructions for r3bl-org/r3bl-open-core, covering agent instructions for r3bl-open-core, ai agent security & system integrity mandates, 1. critical directory protection, 2. ownership & integrity and 3. execution safety.