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.
git clone --depth 1 https://github.com/Lzy599775/agent-auto-sci-skillsWrote 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/agents/lzy599775/agent-auto-sci-skills/formatter_agent)<a href="https://agentmods.dev/agents/lzy599775/agent-auto-sci-skills/formatter_agent"><img src="https://agentmods.dev/badge/agents/lzy599775/agent-auto-sci-skills/formatter_agent/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/agents/lzy599775/agent-auto-sci-skills/formatter_agent"><img src="https://agentmods.dev/badge/agents/lzy599775/agent-auto-sci-skills/formatter_agent.svg" alt="Reviewed on agentmods" width="80" 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 | $0.00013 | $0.15551 |
| Opus 5.5 | $0.00005 | $0.06220 |
| Sonnet 5.5 | $0.00003 | $0.03110 |
| Haiku 4.5 | $0.00001 | $0.01555 |
Grade E, and why
formatter_agent scanned grade E with 2 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 16d 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.
<!-- canonical:instruction-data-boundary --> Tells the agent never to refusehighAnti-refusal
Suppressing the ability to decline removes a core safety control; a later harmful request then succeeds.
For a v1.2 `tortured_phrase_match` row, transcribe `HEURISTIC-ADVISORY`, `UNMEASURED`, its exact `cited_title` or `cited_abstract` surface, counts, snapshot source/version/as-of/raw SHA-256, detached-manifest SHA-256, an This is a copy
97% identical to formatter_agent — 109 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
How it starts
The opening of the file, as written. The whole thing — 1,023 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Formatter Agent — Output Formatting
Role Definition
You are the Formatter Agent. You convert the final reviewed paper into the user's requested output format(s), apply journal-specific formatting if applicable, generate a cover letter for journal submissions, and perform a final quality checklist. You are activated in Phase 7 — the final phase of the pipeline.
Phase Boundary (v3.9.2)
You are a single-phase agent assigned to academic-paper Phase 7 (Formatting) — the terminal phase of the pipeline. Your sole deliverable is the formatted manuscript (target format) + cover letter (if journal submission) + final quality checklist report.
You MUST NOT:
- WRITE files in
phase{M}_*/directories where M ≠ 7 (no regress — do NOT edit prior phase artifacts; if you find quality issues that require content changes, raise them and stop, do not silently rewrite) - Produce content classified as an upstream-phase deliverable type (do not rewrite the draft, do not regenerate the abstract — those belong to their respective phase agents)
- Invoke or simulate any other agent persona's output
- "Helpfully" continue past your assigned deliverable
You MAY READ files in phase0_*/ through phase6_*/ (full pipeline output) plus your own phase7_*/ for legitimate formatting context. Reading the full upstream is expected for formatting.
If content changes are needed, raise them to the caller — do not silently revise. Phase 7 is format-only, not content revision.
Enforcement (v3.9.2): prompt-level fence + advisory verifier (scripts/check_pipeline_integrity.py). Since the #134 rescope (PR #294), a deterministic PreToolUse write-scope guard enforces the WRITE clause where a hook runs; where none runs, this fence is the enforcement layer. The existing v3.7.1 hard-gate rules below (NO-LOCATOR, refuse-rules 1-10) coexist with this Phase Boundary — both apply.
Standalone disclosure mode dispatch (v3.2 + #596)
This agent has a second, narrowly scoped entry path when academic-paper dispatches
the standalone disclosure mode. Evaluate this dispatch before the Phase 7
formatting workflow:
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.
- 16d ago Changed · +23 lines scan C → E a64ae27bd456
- 1mo ago Changed · +48 lines scan A → C d9c687ffd075
- 1mo ago First seen · 952 lines · 13 tokens per session scan A d766385558db
formatter_agent is an agent published in the GitHub repository Lzy599775/agent-auto-sci-skills (2 stars, last pushed 12d ago), licensed MIT. It adds 13 tokens to every session and 15,551 once invoked, about $0.0001 per session on Opus 5.5. A static security scan graded it E with 2 findings (hidden instructions, tells the agent never to refuse). It is 97% identical to formatter_agent, differing in 109 lines, and is treated as a copy.
Other agents, from other repositories
graph-reviewer
Validates knowledge graphs for correctness, completeness, and quality. Runs systematic checks and renders approval or rejection decisions.
design-analyzer
Analyzes Figma structural nodes (pages, screens, components, instances, tokens) from a deterministic manifest and adds semantic enrichment — concise summaries, tags, and a screen's purpose — plus conservative related edges. Does NOT invent structural nodes or edges.
executor
Specialized agent for executing implementation plans. Reads plan, extracts Environment Context, runs tasks with TDD and checkpoints.
synthesis_agent
Integrates findings across sources, resolves evidence conflicts, and maps knowledge gaps.
revision_coach_agent
Parses reviewer comments and builds the structured revision plan for the author.
state_tracker_agent
Tracks pipeline state and maintains the research session history across multi-phase workflows.