Getting it into your agent
This one installs as part of its plugin. Adding the marketplace and installing the plugin brings it with everything else the plugin ships.
/plugin marketplace add ma-ziwei/fathom-mode-for-claude/plugin install fathomWrote 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/commands/ma-ziwei/fathom-mode-for-claude/start)<a href="https://agentmods.dev/commands/ma-ziwei/fathom-mode-for-claude/start"><img src="https://agentmods.dev/badge/commands/ma-ziwei/fathom-mode-for-claude/start.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 5.1 | $0.00025 | $0.00432 |
| Opus 5 | $0.00013 | $0.00216 |
| Sonnet 5 | $0.00005 | $0.00086 |
| Haiku 4.5 | $0.00003 | $0.00043 |
Grade A, and why
start 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 8d 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.
What it actually says
Enter Fathom Mode
User's task: $ARGUMENTS
The UserPromptSubmit hook has already run init_session.py --task "$ARGUMENTS" for this task before this body reached you — a fresh session is live and any prior session has been overwritten. The hook also injected a reminder above this prompt with the exact response format. Follow that reminder. Do NOT call init_session.py yourself.
Defensive fallback (only if hook subprocess silently failed)
Look for a system reminder above starting with [Fathom Mode: new session initialized for task ...]. If it is missing AND the state file shows no active session for this task, the hook subprocess may have failed. In that case, run via Bash using a heredoc with quoted delimiter so apostrophes / em-dashes / dollar signs in $ARGUMENTS pass through literally (no shell escaping needed):
python3 ${CLAUDE_PLUGIN_ROOT}/scripts/init_session.py <<'FATHOM_INIT_END'
{"task": "<JSON-encoded $ARGUMENTS>"}
FATHOM_INIT_END
The closing FATHOM_INIT_END must be at column 0 with no leading whitespace. JSON-encode $ARGUMENTS per JSON rules (escape \" \\ \n) so the payload remains valid regardless of what the user typed.
Then proceed with first-turn extraction + three-part response + Score block per SKILL.md.
Empty $ARGUMENTS (shouldn't happen — bare form is hook-blocked)
If $ARGUMENTS is empty when this body runs, respond with this exact text and stop:
Fathom Mode is ready. To start a session, run
/fathom:start <your task>.
Do NOT call any other tool.
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.
- 8d ago First seen · 34 lines · 25 tokens per session scan A 0569ab3b7277
start is a command published in the GitHub repository ma-ziwei/fathom-mode-for-claude (13 stars, last pushed 4mo ago), licensed MIT. It adds 25 tokens to every session and 432 once invoked, about $0.0001 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 commands, from other repositories
discuss
5-phase requirements-elicitation interview that produces an EARS-validated context/ .md.
sdd
Generate a Kiro-style spec triple — specs/ /{requirements,design,tasks}.md.
drive
Drive a goal to completion over the deterministic firewall — the autonomous goal-driven loop.
brief
Turn a raw request plus attached documents into a validated one-page brief under briefs/ /.
plan
Create a detailed work plan with DoD and save it in memory-bank.
config
Manage execution pipeline.yaml — init / show / validate / path.