agentleFS
tradingstrategy-aitradingstrategy-ai/web3-ethereum-defi/.claude/skills/add-vault-note/SKILL.md

add-vault-note

Skill834 starsChanged 3 days agoMITView on GitHub

Add a note to specific vault

The file

We split this skill into the rules it gives an agent

10 rules. About 0 are common, 7 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈7

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Read and follow the shared description writing style before writing or editing public vault notes or handwritten descriptions.

    line 10
  2. Distinctive:

    Before starting, gather the following information from the user:

    line 15
  3. Distinctive:

    Use search to find the record

    line 26
  4. Distinctive:

    Don't try to open the JSON file, as it is too large

    line 28
  5. Very specific:

    Add the vault address VAULT_FLAGS_AND_NOTES

    line 36
  6. Distinctive:

    If the address already exist ask for user input

    line 38
  7. Distinctive:

    Make sure the address is lowercased

    line 39
  8. Very specific:

    If the user explicitly did not tell you to use any flag, set flag to None

    line 42
  9. Very specific:

    If the user is unsure whether the vault should be flagged, or AI-assisted research does not reach the same conclusion every time, use …

    line 43
  10. Distinctive:

    Run ruff to make sure Python code is formatted.

    line 70

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Add vault note

This skill adds a human-readable note to a specific vault in the database.

Read and follow the shared description writing style before writing or editing public vault notes or handwritten descriptions.

Required inputs

Before starting, gather the following information from the user:

  1. Vault smart contract address - The address of an example vault contract on a blockchain
  2. Message: A human readable message
  3. Flags (optional): Flags from VaultFlag enum - if not given set to None

1. Resolve vault name

Can be resolved from https://top-defi-vaults.tradingstrategy.ai/top_vaults_by_chain.json JSON file

  • It's a list of JSON object records
  • Use search to find the record
  • Extract its name
  • Don't try to open the JSON file, as it is too large

Alternatively use check-vault-onchain.py.

2. Edit the flag.py

The vault notes are stored in ``eth_defi/vault/flag.py`.

Add the vault address VAULT_FLAGS_AND_NOTES

  • If the address already exist ask for user input
  • Make sure the address is lowercased
  • For the message create a Python constant like other messages have in flag.py
  • Set the message and flag in the dictionary
  • If the user explicitly did not tell you to use any flag, set flag to None
  • If the user is unsure whether the vault should be flagged, or AI-assisted research does not reach the same conclusion every time, use VaultFlag.review_needed. It does not blacklist the vault; its note tells readers the vault is under review. For a review a person asks for, use the REVIEW_NEEDED_MANUAL message constant. The check-specific constants REVIEW_NEEDED_OFF_MARKET_COLLATERAL, REVIEW_NEEDED_EXIT_LIQUIDITY and REVIEW_NEEDED_DATA_QUALITY say the automated investability checks did not agree, so use them only for findings of the vault report investability check. Record in the comment block what is uncertain and what a reviewer should check to decide
  • Add an extensive line comment above the entry, as the Instructions for adding entries in the flag.py module docstring require: the vault name, protocol and chain, the date, what is wrong and why this flag, and canonical source URLs (the tradingstrategy.ai vault page, the block explorer page, and any announcement or incident report). A bare address or a name-only comment is not enough

Persisted metadata refresh

New or changed flags require a numbered refresh of existing rows: export reads the stored _flags and does not recompute them from the manual table. Only note-only changes can rely on the current manual notes being resolved during export; document when that makes a migration unnecessary. Cached descriptions and other persisted fields also need a numbered repair when they must be refreshed. Follow the migration documentation and use the reviewed chain/address and fields in a new script. Standalone fix-*.py and migrate-*.py repairs are legacy format and must not be created or run for this task. Use the unified runner and include its production run reminder; do not mutate the local database during authoring.

3. Format code

Run ruff to make sure Python code is formatted.

4. Open PR

Finish with clean-up-for-pr at the outermost task's review handoff.

  • Open a pull request.
  • This is not a feature pull request - no need to add changelog entry
  • Prefix the pull request "notes" instead of "feat"

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.