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/marcus/td/claude-mdgit clone --depth 1 https://github.com/marcus/tdWhat 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 | $0.01611 | $0.01611 |
| Opus 5 | $0.00805 | $0.00805 |
| Sonnet 5 | $0.00322 | $0.00322 |
| Haiku 4.5 | $0.00161 | $0.00161 |
Grade C, and why
td CLAUDE.md scanned grade C with 1 finding 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 2d 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.
Hidden instructionshighPrompt injection
Directives inside HTML comments, invisible characters or bidirectional overrides are read by the model and not by the person reviewing the file.
<!-- td-agent-instructions:start --> How it starts
The opening of the file, as written. The whole thing — 126 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
Working with td
td keeps task context durable across sessions. In a new context, run td usage --new-session -q to see current work.
Use your judgment about how much tracking a task needs. For substantive work: td start <id>, record progress with td log, hand off with td handoff <id>, then td review <id>.
Closing needs a review. Say who did it (default trusted mode; delegated/strict allow only the first):
- independent session:
td approve <id> --reason "..." - a sub-agent:
td approve <id> --reviewed-by "<who>" - you:
td approve <id> --self-review --reason "..."
Prefer a reviewer with its own TD_CONTEXT_ID; never name one who did not review.
Run td usage or td <command> --help.
Review Model (Trusted Review)
Work still needs a review before it closes. What changed is that td now asks you to say who reviewed it, rather than assuming the reviewing agent and the recording session are the same thing.
In the default trusted mode:
- An independent session reviews and closes:
td approve <id> --reason "..." - A reviewer attests without closing and anyone closes after:
td approve <id> --record-only --reason "...", thentd approve <id> --reason "..." - You implemented it and a sub-agent reviewed it:
td approve <id> --reviewed-by "<who>" - You implemented it and reviewed it yourself:
td approve <id> --self-review --reason "..."
Prefer the first two. When a reviewing sub-agent has its own TD_CONTEXT_ID its
independence is mechanically verified rather than asserted — see
docs/multi-agent-sessions.md.
--reviewed-by is an attestation td cannot verify. It exists so an orchestrator
recording a sub-agent's review writes a true record instead of claiming a
self-review it did not perform. Do not use it to launder an unreviewed change:
naming a reviewer who did not review is worse than --self-review, because it
reads as independent in the audit trail. Do not create a throwaway session to
make a review appear independent either.
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.
- 2d ago First seen · 126 lines · 1,611 tokens per session scan C a3b2deabdd62
td CLAUDE.md is an instructions file published in the GitHub repository marcus/td (244 stars, last pushed 5d ago), licensed MIT. It adds 1,611 tokens to every session, about $0.0081 per session on Opus 5. A static security scan graded it C with 1 finding (hidden instructions). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
pi-fabric AGENTS.md
Instructions for monotykamary/pi-fabric, covering agents.md, golden rule: build when done, before committing and commits.
opendream AGENTS.md
Instructions for pylit-ai/opendream, covering agents.md, release hygiene, path-scoped guidance and opendream activation.
octowiz CLAUDE.md
Instructions for raelli/octowiz, covering octowiz, releasing & deploying, commands, agent skills and issue tracker.
trellis AGENTS.md
Instructions for terraboops/trellis, covering trellis agent guidelines, before every commit, after changing pipeline logic and release checklist.
hatch3r CLAUDE.md
Instructions for hatch3r/hatch3r, covering hatch3r — development instructions, architecture, development commands, two-axis pillar framework (2.0.0) and orchestrator self-discipline (bypass protection).
ai-usagebar CLAUDE.md
Instructions for akitaonrails/ai-usagebar, covering claude.md, release checklist — must do all of these, hard invariants — never break these, secret-discipline rules (learned the hard way) and live api smoke discipline.