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 agents/metedata/remarc/mcp-tools-referencegit clone --depth 1 https://github.com/metedata/RemarcWhat 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 | $0.00022 | $0.00516 |
| Opus 5 | $0.00011 | $0.00258 |
| Sonnet 5 | $0.00004 | $0.00103 |
| Haiku 4.5 | $0.00002 | $0.00052 |
Grade A, and why
mcp-tools-reference 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 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.
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 — 40 lines — stays where its author put it; the contents beside it link to each section on GitHub.
The Remarc MCP (Model Context Protocol) server exposes seven tools. Every connected agent gets the same set regardless of how it was installed; setup paths are in the agent overview.
Tools
Each tool and what it does:
| Tool | What it does |
|---|---|
remarc_list_sessions |
Lists active sessions, marks the active one, and shows open, in-progress, and resolved counts. |
remarc_list_comments |
Lists comments, optionally filtered by session_id, status, or type. Types: comment, screenshot, quickNote, critMode, webElement. |
remarc_get_comment |
Returns one comment by full UUID or the 5-character short ID shown on the card. Includes the image path for screenshot comments and the CSS selector, component name, and file path for web element comments. |
remarc_set_status |
Sets one comment's status. resolved requires a summary of what was done, which appears on the card in Remarc. |
remarc_bulk_set_status |
Sets the status of many comments at once, with a shared summary or per-comment summaries. |
remarc_rename_session |
Renames a session. |
remarc_create_session |
Creates a session paired to the current agent conversation and makes it active. |
Statuses
Agents use these exact status strings:
| Status | Meaning |
|---|---|
open |
Not yet accepted by the agent. |
handedOff |
You handed the comment to the agent to address. |
inProgress |
The agent is actively working this comment. |
resolved |
Done, with a resolution summary. |
These map to the Open, Handed Off, In-Progress, and Resolved dots on comment cards; see statuses and history.
Deliberate limits
No tool deletes or closes sessions. Session cleanup happens in the app (or through your integration's conversation-cleared setting), so an agent can never destroy a container of your feedback.
Agents should not edit comments.json directly. The MCP tools are the only safe write path; hand-editing the data file risks losing comments to concurrent writes.
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 · 40 lines · 22 tokens per session scan A 064495c2f248
mcp-tools-reference is an agent published in the GitHub repository metedata/Remarc (66 stars, last pushed 11d ago), licensed MIT. It adds 22 tokens to every session and 516 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 agents, from other repositories
swift-mcp-expert
Expert assistance for building MCP servers in Swift with the FastMCP library. Use proactively when authoring @Tool / @MCPResource / @MCPPrompt types, configuring the FastMCP builder, defining @Generable schemas, or debugging FastMCP server issues.
i18n
Multiplex ships English (source), Traditional Chinese (zh-Hant, Taiwan wording) and Japanese (ja). Decisions below were settled 2026-08-18; do not re-litigate without new facts.
Native SwiftUI Builder
Builds complex custom SwiftUI views using standard SwiftUI components with TTBaseUIKit design tokens (XView/XSize/XFont). Does NOT use TTBaseSUI wrapper components.
kmp-architect
KMP architecture expert. Designs shared module structures, expect/actual patterns, navigation, and DI for Kotlin Multiplatform projects.
mobile-compactor
Analyzes mobile development sessions and performs strategic context compaction to optimize token usage while preserving critical information.
backends-and-probe
Load-bearing decisions split from AGENTS.md — read before touching TmuxProbe, HerdrProbe, session creation/targeting, or file attach/drop.