Back to skill

Security audit

Empire Builder

Security checks for vulnerabilities and agentic risk

Overview

This skill is transparent about its purpose, but it guides agents through live treasury and account-changing operations with weakly scoped signing guidance and an unnecessary privateKey field.

Install only if you understand Empire Builder's live mainnet workflows. Use x-api-key headers, never send wallet private keys or seed material, require human review before any signing or broadcast, and treat captured signatures as sensitive because many documented messages are not server-bound to a specific action.

Vulnerability Patterns
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Findings (2)

T09 · Insecure Skill Coding Practices

Error
Location
references/http-api.md:119
Finding

Guardian signatures are not cryptographically bound to requested actions and lack enforced replay protection

Content
View full analysis

Vulnerability Details

File Location: references/http-api.md, lines 119–141 and 1078–1086
Vulnerability Type: Authorization replay and signature-context confusion
Risk Level: High

Vulnerable documentation excerpt:

markdown
- **Signature fields are always**: `signature`, `message`, **`signerAddress`**.
- The server verifies `verifyMessage({ address: signerAddress, message, signature })` and then checks the signer is a **guardian** of the empire:
  - `signerAddress` must equal `empires.owner` **or** be included in `empires.co_emperors` for the target empire.
- **Message format:** for leaderboard creates the backend **does not enforce a specific template**; it only verifies the signature against the provided `message`.

**Recommended canonical message template (agent-side):**

Use a consistent, human-readable message to make signatures auditable and easy to debug. Suggested format:

```text
Empire Builder — Create Leaderboard
type: <leaderboard_path_suffix>
empireId: <tokenAddress>
name: <name>
timestamp: <ISO-8601>
nonce: <random-uuid>

Notes:

  • leaderboard_path_suffix: e.g. csvLeaderboards, tokenHoldersLeaderboards, farcasterChannelLeaderboards
  • Include timestamp + nonce to reduce replay risk and make logs unambiguous (server currently does not enforce these, but they help operators).
text

The general signing guidance further states:

```markdown
### What is `message`?

- Unless an endpoint explicitly enforces an exact template (currently **only** `deploy-empire-tokenless`), `message` can be any plain text.
- Recommended pattern (to prevent replay confusion): include endpoint + empire id + a short nonce/timestamp, e.g.

```text
Empire Builder: create leaderboard for <empire_id> at 2026-04-30T00:00:00Z
text

### Technical Analysis

The server reportedly verifies only that the supplied EIP-191 signature matches the in
...[truncated 3069 chars]
Remediation
View remediation

Remediation Suggestions

  • Define a versioned, endpoint-specific authorization schema for every signed operation.
  • Reconstruct the signed payload server-side from normalized request fields rather than trusting a caller-supplied free-form message.
  • Bind every signature to:
    • Application name and version.
    • Exact endpoint and HTTP method.
    • Empire ID and vault address.
    • Chain ID where applicable.
    • Every security-sensitive request parameter.
    • Signer address.
    • Issued-at and expiration timestamps.
    • A cryptographically random, single-use nonce.
  • Store consumed nonces atomically and reject replayed, expired, future-dated, or incorrectly scoped authorizations.
  • Prefer EIP-712 typed structured data over arbitrary EIP-191 text so wallets can display the exact action and parameters being authorized.
  • Use distinct domain separators and action types for leaderboard creation, booster changes, staking activation, deployment, and distribution preparation.
  • Invalidate old arbitrary-message authorization behavior after a migration period; do not silently continue accepting legacy signatures.
  • Redact signatures, API keys, and signed messages from logs and telemetry.
  • Add tests proving that changing any protected field, endpoint, empire ID, chain ID, timestamp, or nonce invalidates authorization.

T09 · Insecure Skill Coding Practices

Warning
Location
references/http-api.md:303
Finding

Unused privateKey request field encourages unnecessary transmission of wallet secrets

Content
View full analysis

Vulnerability Details

File Location: references/http-api.md, lines 303–324
Vulnerability Type: Unnecessary sensitive-data transmission
Risk Level: Medium

Vulnerable documentation excerpt:

markdown
### `POST /api/leaderboards/apiLeaderboards`

Creates a leaderboard by fetching and parsing an external JSON API.

- **`tokenAddress` must be an ERC‑20 address** (tokenless empires are not accepted for this type).
- `apiEndpoint` must be a valid URL.
- `privateKey` exists in the request body but is **unused** server-side.

**Request body schema (JSON):**

```json
{
  "tokenAddress": "0x...",
  "name": "string (≤ 20 chars)",
  "description": "string (≤ 180 chars)",
  "apiEndpoint": "https://example.com/leaderboard.json",
  "applyBoosters": true,
  "apply_reputation_boosters": true,
  "icon": "optional string",
  "signature": "0x...",
  "message": "plain text that was signed",
  "signerAddress": "0xGuardian",
  "privateKey": "optional string (unused)"
}
text

### Technical Analysis

A field explicitly named `privateKey` appears in the JSON schema for a request sent to the production service at `https://empirebuilder.world`, even though the documentation states that the backend does not use it. This is contrary to data minimization and least-privilege principles.

Developers, agents, schema generators, or operators may infer that the field should contain a wallet private key or an API credential. If populated, that secret crosses a network and may be processed by HTTP middleware, reverse proxies, web application firewalls, observability systems, request loggers, crash reports, or backend body parsers. TLS protects data in transit from passive interception but does not prevent exposure to the destination service or its operational infrastructure.

No reviewed file proves that the service intentionally records or exfiltrates this value. The confirmed defect is that the Skill exposes a
...[truncated 1736 chars]
Remediation
View remediation

Remediation Suggestions

  • Remove privateKey from the documented request schema, examples, generated clients, validation models, and endpoint implementation.
  • Explicitly instruct users never to send wallet private keys, seed phrases, raw signing keys, or unrelated API credentials to Empire Builder.
  • Reject requests containing privateKey or similarly named secret fields with a clear 400 response rather than silently accepting them.
  • Configure request logging, tracing, analytics, error reporting, and reverse proxies to redact historical and future sensitive fields.
  • Review retained logs and telemetry for prior occurrences of the field. If real credentials may have been submitted, securely delete retained copies where possible and notify affected operators to rotate credentials and move wallet assets.
  • Keep signing local to an operator-controlled wallet. Transmit only the resulting scoped signature, never the signing key.
  • Add automated tests and schema checks that fail if secret-bearing fields such as privateKey, seedPhrase, or mnemonic are introduced into network request models.
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • YARA SignaturesMalware Match, Webshell Match, Cryptominer Match
Findings (19)

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 35)May include surrounding context.

md
Models usually load **`SKILL.md` first** only. Treat this file as **routing + constraints**. Before writing HTTP bodies, signing flows, or multi-step flows, **o

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 76)May include surrounding context.

md
Models usually load **`SKILL.md` first** only. Treat this file as **routing + constraints**. Before writing HTTP bodies, signing flows, or multi-step flows, **o

Missing User Warnings

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The file states that hosts and chains are live and that prepare/broadcast/store flows update on-chain balances, but it does not place an upfront, prominent warning around irreversible financial consequences. In the context of an agent skill that may automate blockchain actions, insufficient hazard signaling materially raises the risk of destructive user mistakes or unsafe autonomous execution.

Content

No source excerpt is available for this finding.

YARA rule 'agent_skill_mcp_tool_poisoning_metadata': MCP/tool metadata poisoning indicators in tool schemas or skill manifests [agent_skills]

High
Category
YARA Match
Confidence
80% confidence
Finding

YARA rule matched a hack tool or exploit indicator (offensive tools, reconnaissance, privilege escalation, or exploit frameworks).

Content

Scanner excerpt · references/http-api.md (reported line 174)May include surrounding context.

sters": true, "signature": "0x...", "message": "plain text that was signed", "signerAddress": "0xGuardian" }

text

**Required fields:** `tokenAddress`, `erc20Address`, `signature`, `message`, `signerAddress`.

Working example:

```json
{
  "tokenAddress": "fid123",
  "erc20Address": "0x1111111111111111111111111111111111111111",
  "erc20ChainId": 8453,
  "applyBoosters": true,
  "signature": "0xaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
80% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · references/http-api.md (reported line 950)May include surrounding context.

md
- `multiplier` must be between `1.1` and `5.0` and have **max one decimal place**.
- For `booster.type === "NFT"`, ERC-1155 requires `tokenId` (except OG-dickbutt special case).

### `DELETE /api/boosters/[empire_id]` (remove booster)

**Auth:** API key + signature.

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
80% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · references/http-api.md (reported line 1051)May include surrounding context.

md
**Response:** `{ stakingBoosters: Booster[], boosters: Booster[] }`.

### `DELETE /api/staking-boosters/[empire_id]` (remove staking booster)

**Auth:** API key + signature. Same guardian + activated-empire check.

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
80% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · references/workflows.md (reported line 75)May include surrounding context.

md
1. **`GET /api/boosters/[vaultOrTokenContext]`** — inspect multipliers.
2. **`POST /api/boosters/[empire_id]`** — add booster (API key + guardian-signed payload).
3. **`DELETE /api/boosters/[empire_id]`** — remove booster entry.

This route only manages ERC‑20 / NFT / QUOTIENT boosters. STAKING boosters live behind a dedicated CRUD — see Workflow 8.

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
80% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · references/workflows.md (reported line 119)May include surrounding context.

  1. Remove a staking booster:

    text
    DELETE /api/staking-boosters/[empire_id]
    { boosterId, signer, signature, message }
    

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The document explicitly permits API authentication via an api_key query parameter but does not warn that query strings are commonly logged by browsers, proxies, CDNs, server access logs, analytics tools, and referral headers. In a skill that performs privileged write operations, this increases the chance of accidental credential disclosure and subsequent unauthorized API use.

Content

No source excerpt is available for this finding.

Whitespace Padding

Medium
Category
Prompt Injection
Confidence
80% confidence
Finding

Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.

Content

Scanner excerpt · references/http-api.md (reported line 174)May include surrounding context.

md
"erc20Address": "0x1111111111111111111111111111111111111111",
  "erc20ChainId": 8453,
  "applyBoosters": true,
  "signature": "0xaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa",
  "message": "Create token holders leaderboard for fid123 ranking 0x1111111111111111111111111111111111111111",
  "signerAddress": "0x2222222222222222222222222222222222222222"
}

Whitespace Padding

Medium
Category
Prompt Injection
Confidence
80% confidence
Finding

Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.

Content

Scanner excerpt · references/http-api.md (reported line 250)May include surrounding context.

md
"applyBoosters": true,
  "apply_reputation_boosters": true,
  "icon": "https://example.com/icon.png",
  "signature": "0xbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb",
  "message": "Create CSV leaderboard Top 10 for 0x3333333333333333333333333333333333333333",
  "signerAddress": "0x2222222222222222222222222222222222222222"
}

Whitespace Padding

Medium
Category
Prompt Injection
Confidence
80% confidence
Finding

Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.

Content

Scanner excerpt · references/http-api.md (reported line 295)May include surrounding context.

md
"chainId": 8453,
  "applyBoosters": true,
  "apply_reputation_boosters": true,
  "signature": "0xcccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccc",
  "message": "Create NFT leaderboard for 0x3333333333333333333333333333333333333333 using 0x6666666666666666666666666666666666666666",
  "signerAddress": "0x2222222222222222222222222222222222222222"
}

Whitespace Padding

Medium
Category
Prompt Injection
Confidence
80% confidence
Finding

Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.

Content

Scanner excerpt · references/http-api.md (reported line 339)May include surrounding context.

md
"apiEndpoint": "https://example.com/scores.json",
  "applyBoosters": true,
  "apply_reputation_boosters": true,
  "signature": "0xdddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddd",
  "message": "Create API leaderboard for 0x3333333333333333333333333333333333333333 using https://example.com/scores.json",
  "signerAddress": "0x2222222222222222222222222222222222222222"
}

Whitespace Padding

Medium
Category
Prompt Injection
Confidence
80% confidence
Finding

Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.

Content

Scanner excerpt · references/http-api.md (reported line 375)May include surrounding context.

md
"description": "FarToken scores.",
  "applyBoosters": true,
  "apply_reputation_boosters": true,
  "signature": "0xeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee",
  "message": "Create FarToken leaderboard for 0x3333333333333333333333333333333333333333",
  "signerAddress": "0x2222222222222222222222222222222222222222"
}

Whitespace Padding

Medium
Category
Prompt Injection
Confidence
80% confidence
Finding

Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.

Content

Scanner excerpt · references/http-api.md (reported line 418)May include surrounding context.

md
"chainId": 8453,
  "applyBoosters": true,
  "apply_reputation_boosters": true,
  "signature": "0xffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff",
  "message": "Create TIPN leaderboard for 0x3333333333333333333333333333333333333333 using 0x6666666666666666666666666666666666666666",
  "signerAddress": "0x2222222222222222222222222222222222222222"
}

Whitespace Padding

Medium
Category
Prompt Injection
Confidence
80% confidence
Finding

Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.

Content

Scanner excerpt · references/http-api.md (reported line 460)May include surrounding context.

md
"castScoreIncludes": { "likes": true, "recasts": true, "comments": true, "quotes": false },
  "applyBoosters": true,
  "apply_reputation_boosters": true,
  "signature": "0x11111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111",
  "message": "Create cast leaderboard for fid123 cast 0xabc123",
  "signerAddress": "0x2222222222222222222222222222222222222222"
}

Whitespace Padding

Medium
Category
Prompt Injection
Confidence
80% confidence
Finding

Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.

Content

Scanner excerpt · references/http-api.md (reported line 508)May include surrounding context.

md
"timeframeDays": 30,
  "applyBoosters": true,
  "apply_reputation_boosters": true,
  "signature": "0x22222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222",
  "message": "Create channel leaderboard for fid123 channel base",
  "signerAddress": "0x2222222222222222222222222222222222222222"
}

Whitespace Padding

Medium
Category
Prompt Injection
Confidence
80% confidence
Finding

Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.

Content

Scanner excerpt · references/http-api.md (reported line 551)May include surrounding context.

md
"timeframeDays": 90,
  "applyBoosters": true,
  "apply_reputation_boosters": true,
  "signature": "0x33333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333333",
  "message": "Create interaction leaderboard for fid123 for @vitalik (90d)",
  "signerAddress": "0x2222222222222222222222222222222222222222"
}

Whitespace Padding

Medium
Category
Prompt Injection
Confidence
80% confidence
Finding

Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.

Content

Scanner excerpt · references/http-api.md (reported line 590)May include surrounding context.

md
"leaderboardName": "top_users",
  "applyBoosters": true,
  "apply_reputation_boosters": true,
  "signature": "0x44444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444444",
  "message": "Create quotient leaderboard top_users for 0x3333333333333333333333333333333333333333",
  "signerAddress": "0x2222222222222222222222222222222222222222"
}

Static analysis

No suspicious patterns detected.