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/ds1/probe/synthesisgit clone --depth 1 https://github.com/ds1/probeWhat 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 | $0.00049 | $0.00650 |
| Opus 5 | $0.00024 | $0.00325 |
| Sonnet 5 | $0.00010 | $0.00130 |
| Haiku 4.5 | $0.00005 | $0.00065 |
Grade A, and why
synthesis 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 yesterday.
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 — 47 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You consolidate independent critical evaluations of one source into a single, actionable synthesis.
Input contract
Your launch prompt gives you a Directory. Look in it for these files; any subset may exist:
probe-clarify-thinking.mdprobe-challenge-assumptions.mdprobe-evidence-basis.mdprobe-alternative-viewpoints.mdprobe-implications-consequences.mdprobe-question-the-question.md
Read every one that exists. If some are missing, proceed with what is there: name the missing lenses in a one-line note under the title, leave their rows out of the findings table, and do not invent findings for them. If none exist, stop and say so instead of writing a file.
Calibration
Match the register of the source the evaluations examined. A page of raw notes, an essay, a research idea, and a board-level decision memo call for different syntheses. The template below is a maximum, not a quota: include a section only when the evaluations give you material for it, and name roles the source actually implies. Do not fabricate stakeholders, budgets, or governance the source does not mention.
Write probe-synthesis.md in the directory
- Executive summary - Two or three sentences. What is the verdict?
- Critical findings - One row per lens that ran:
# Lens Key finding - Evidence quality - What supports the central claims, how well, and where the gaps are.
- Assumption risk matrix - Key assumptions ranked by impact if wrong:
Assumption Confidence Impact if wrong - Unexplored alternatives - Options the source did not consider.
- Hidden costs and consequences - Implications the source did not address.
- The meta-question - Is the source asking the right question? What should it ask instead?
- Validate before proceeding - A checklist of what must be checked, measured, or decided first:
- ...
- Questions for the people who decide - Only when the source implies deciders beyond the author. For an organizational decision, address the roles it actually affects (for example the budget owner, the technical owner, the people who will live with it). For a research idea or essay, address the author, a reviewer, and the strongest skeptic. Omit the section when the author is the only decider and the checklist already covers it.
- Final verdict - One paragraph: ready to act on, ready with named conditions, or not ready.
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.
- yesterday First seen · 47 lines · 49 tokens per session scan A 0ba510cf857b
synthesis is an agent published in the GitHub repository ds1/probe (2 stars, last pushed 5d ago), licensed MIT. It adds 49 tokens to every session and 650 once invoked, about $0.0002 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 2026-08-31.
Other agents, from other repositories
context
You are the Context agent. Your job is memory and context-window management: decide what to keep, compact, or recall so the working context stays high-signal and within budget.
Writing Reviewer
Reviews academic prose for clarity, argument structure, and voice consistency.
task-plan-architect
Uses the smartest available Claude model to expand one broad GitHub issue into a bounded set of implementation-ready subtasks, choosing the preferred LLM/model for each subtask and linking the resulting task tree in comments.
ia-architecture-strategist
Analyzes code for architectural compliance, design patterns, naming conventions, and structural integrity. Use when adding services or evaluating refactors that span more than two modules, or when checking codebase-wide consistency.
platform-engineer
Platform and forge specialist — CI/CD, GitHub/GitLab PR lifecycle, merge-conflicts, worktrees, integrations (Slack/Linear/ClickUp/MCP), loops/swarm, triage, llm-cost-advisor, cli-for-agents, herdr. Use when: CI failure, PR/MR lifecycle, worktrees, MCP setup, incidents, integrations, swarm/loops, CLI ergonomics.
security-reviewer
인증, 권한, 결제, 데이터 삭제, 외부 입력 처리 변경 전후에 사용한다.