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 skills add TeresaPeng-zju/be-water --skill bewater-business-memorygit clone --depth 1 https://github.com/TeresaPeng-zju/be-waterWrote 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/skills/teresapeng-zju/be-water/bewater-business-memory)<a href="https://agentmods.dev/skills/teresapeng-zju/be-water/bewater-business-memory"><img src="https://agentmods.dev/badge/skills/teresapeng-zju/be-water/bewater-business-memory/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/skills/teresapeng-zju/be-water/bewater-business-memory"><img src="https://agentmods.dev/badge/skills/teresapeng-zju/be-water/bewater-business-memory.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 | $0.00133 | $0.00699 |
| Opus 5 | $0.00067 | $0.00349 |
| Sonnet 5 | $0.00027 | $0.00140 |
| Haiku 4.5 | $0.00013 | $0.00070 |
Grade A, and why
bewater-business-memory 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 12d 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 — 52 lines — stays where its author put it; the contents beside it link to each section on GitHub.
BeWater Business Memory
Treat business memory as an evidence system, not a generic summary or advice generator.
Core workflow
- Locate the user's raw records. Accept pasted text or files such as Markdown, text, JSON, and CSV.
- For a directory of files, run
python3 scripts/prepare_records.py <path> --output <workspace.json>to create stable source references. Do not run it for a single short pasted record. - Classify each source as client chat, meeting transcript, delivery note, customer feedback, manual note, or other.
- Extract only claims supported by the source. Separate facts, business events, outcome claims, and unknowns.
- Propose case-state changes independently across commercial, delivery, payment, and outcome. Never apply an ambiguous proposal as confirmed.
- When two or more independent cases exist, compare them for recurring signals. Call a single occurrence an observation, not a pattern.
- Return a concise business-memory update with source references beside every observation or pattern.
- Suggest a reusable asset only when repeated work or validated material exists. Preserve its source cases.
Read references/evidence-rules.md before interpreting records. Read references/output-model.md when producing structured JSON or updating a persistent workspace.
Evidence discipline
- Preserve the raw material and give each source a stable
source_ref. - Quote only the shortest evidence fragment needed to verify a claim.
- Treat known context as background, never as new evidence.
- Do not infer causation from correlation, silence, or missing fields.
- Distinguish customer self-report, observable change, and independently verified outcome.
- Ask only questions that would materially change identity resolution, event type, or case state.
- Match the user's language.
Response shape
For a single record, return:
发生了什么— a short operating update.可核对事实— facts with evidence and source refs.经营事件— what changed, event time if known, and direct next actions.状态建议— dimension, from/to, evidence, confidence, and whether confirmation is required.仍需确认— only material ambiguities.
What ships with it
4 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 12d ago First seen · 52 lines · 133 tokens per session scan A 66985c650eb9
bewater-business-memory is a skill published in the GitHub repository TeresaPeng-zju/be-water (24 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 133 tokens to every session and 699 once invoked, about $0.0007 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 skills, from other repositories
obsidian-bases
Explain, draft, and validate Obsidian Bases .base files with filters, formulas, properties, summaries, and table, card, or list views. Use for Obsidian Bases, database-like vault views, dynamic tables, reading lists, task trackers, filters, formulas, summaries, and .base file edits.
wiki-retrieve
Build and query a vault-local contextual BM25 retrieval index with optional multilingual Nomic cosine reranking; use for retrieve, hybrid retrieval, BM25, rerank, contextual retrieval, chunk search, vault search, semantic search, find relevant passages, or retrieval diagnostics. Derived caches stay under .vault-meta…
okf
Author, maintain, and consume Open Knowledge Format (OKF) knowledge bundles — portable markdown + YAML frontmatter that both humans and agents read. Use when capturing project knowledge (services, APIs, schemas, metrics, runbooks, decisions) into an OKF bundle, when updating one after code or docs change, or when a…
shodh-memory
Persistent memory system for AI agents. Use this skill to remember context across conversations, recall relevant information, and build long-term knowledge. Activate when you need to store decisions, learnings, errors, or context that should persist beyond the current session.
vault-health-batch
Autonomous batch mode for Vault Health findings. Works through orphans, missing backlinks, and tags in batches without asking on every fix.
memex-sync
Sync Zettelkasten cards across devices via git.