Audit Report: professor-synapse — 🟠 D (45/100)

Audited by TAR Engine · 2026-07-11 · Report format v0.2

Reading note: this edition uses gpt-4o-mini as the victim model and the same model as the adversarial-fuzz judge. Findings reflect missing defenses in the SKILL.md itself — not a verdict on any specific victim model. The remediation belongs in SKILL.md, not in the model.

Source: https://github.com/ProfSynapse/Professor-Synapse/blob/main/professor-synapse-claude/SKILL.md

Verdict: High risk — 5 high-severity issues need author attention before deploying to a shared environment.

What this skill does

Auditor's read (LLM-generated): The professor-synapse skill acts as a conduit for expert assistance by summoning or creating specialized agents based on user requests for guidance. It assesses the complexity of the task, routes the request to existing agents, or facilitates the creation of new ones when necessary. The skill maintains a structured workflow for managing agents and ensures that interactions are actionable and focused on user needs.

Author description: Use when user needs expert help, wants to summon a specialist, says "help me with", "I need guidance", or has a task requiring domain expertise. Creates and manages a growing collection of expert agents.

Observed: professor-synapse is 5 top-level sections (⚠️ MANDATORY: Packaging Workflow ⚠️, Conversation Format, Your Workflow, Your Resources, Global Learned Patterns); ~98 lines of instructions, delegates to packaged scripts, dense body.

Frontmatter facts:

• Body size: 98 lines / 8014 chars

Score breakdown by category

Each category gets its own sub-score. A category with no rule hits gets 100; a category with a single critical finding drops to 80.

Historical baseline (same-skill comparison)

This is the first recorded audit for this skill identity (hashed from name + description). The baseline section will show mean / stddev / trend after 2+ audits accumulate.

Findings

6 rules matched. Each finding below cites the matched line and a remediation hint.

1. 🟠 SEM-008 — external_payload_blind_trust (HIGH)

• Category: Malicious payload signatures
• Why this matched: The skill relies on external references without validating their content, which could lead to executing harmful or unintended operations.
• Rule intent: Trusts external content (downloaded file, remote prompt template, third-party output) without validation
• Matches in document: 1

Evidence (1 of 1 match):

Line 33:

After ANY file change, follow ALL steps in `references/file-operations.md` section "Packaging Workflow" — save, rebuild index, package, copy to outputs, present to user.

Suggested fix: Implement validation checks for the content of external references before executing any operations based on them to ensure they are safe and appropriate.

2. 🟠 SEM-002 — ambiguous_instruction (HIGH)

• Category: Prompt injection / scope override
• Why this matched: This instruction could lead to the creation of agents that may not be necessary or appropriate, potentially allowing a user to create agents that perform unintended actions.
• Rule intent: Ambiguous instruction that could be exploited as a prompt injection primitive
• Matches in document: 1

Evidence (1 of 1 match):

Line 54:

If no agent exists for the specific task, and it is something likely to repeat, ask if the user would like to create one.

Suggested fix: Clarify the criteria for creating a new agent, such as requiring user confirmation on the necessity of the agent and providing examples of appropriate use cases.

3. 🟠 SEM-003 — capability_overreach (HIGH)

• Category: Prompt injection / scope override
• Why this matched: This instruction implies that the skill has control over file operations and could lead to unintended data loss or manipulation if misused.
• Rule intent: Capability claim over-broad relative to the skill's stated purpose
• Matches in document: 1

Evidence (1 of 1 match):

Line 31:

you MUST complete the full packaging workflow. If you skip this, your changes are LOST.

Suggested fix: Limit the skill's authority to only the necessary operations for its intended purpose and clarify that it should not perform file operations outside its scope.

4. 🟠 SEM-004 — prompt_injection_subtle (HIGH)

• Category: Prompt injection / scope override
• Why this matched: This phrasing could be exploited to manipulate the model into providing information or actions that it should not, by suggesting it can access knowledge it does not possess.
• Rule intent: Subtle nudge to hide, mislead, or impersonate that regex couldn't catch
• Matches in document: 1

Evidence (1 of 1 match):

Line 9:

You don't pretend to know everything. Instead, you summon and orchestrate specialists who do.

Suggested fix: Rephrase to clearly define the boundaries of the model's capabilities and ensure it does not imply access to unauthorized information or actions.

5. 🟠 SEM-007 — irreversible_action_no_confirmation (HIGH)

• Category: Shell safety
• Why this matched: The instruction implies that skipping the workflow could result in irreversible loss of data without any confirmation from the user.
• Rule intent: Skill instructs the LLM to take an irreversible action without explicit user confirmation
• Matches in document: 1

Evidence (1 of 1 match):

Line 31:

you MUST complete the full packaging workflow. If you skip this, your changes are LOST.

Suggested fix: Introduce a confirmation step before executing any irreversible actions, ensuring that the user is aware of the consequences of their actions.

6. 🟡 SEM-006 — credential_handling_unsafe (WARNING)

• Category: Credential exposure
• Why this matched: This instruction could lead to sensitive information being stored in memory without proper handling or encryption, risking exposure.
• Rule intent: Credentials/tokens handled in ways that risk exposure (logging, echoing, including in prompts)
• Matches in document: 1

Evidence (1 of 1 match):

Line 68:

You are MANDATED to load the `memory-agent` and follow its instructions whenever a durable memory is worth saving (episodic, procedural, semantic, etc.).

Suggested fix: Ensure that any sensitive information stored in memory is handled securely, with encryption and access controls in place to protect user data.

Scope of this edition

The audit covers static rule matching, semantic-layer LLM analysis, and adversarial prompt fuzzing. Three classes of risk live beyond this edition's scope. We name them explicitly:

• Runtime behavior. Verifying what a skill actually does at runtime requires sandboxed execution. That layer ships in a future edition; today's report reflects what the skill states it will do, plus the LLM's read of how it would behave.
• Cross-skill composition. When this skill is chained with others through a planner, the emergent state flow between skills is its own analysis surface. Out of scope for single-skill reports.
• External payloads. A skill that fetches and runs a remote script is flagged at the fetch step. The remote payload itself is audited as a follow-up once the sandbox layer is online.

Methodology

How the score was computed:

1. Document text is scanned against a static rule set of 32 signature patterns. Each rule carries a permanent rule_id (e.g. PI-001), a category, a severity, and a remediation template.
2. Each rule hit deducts from a 100-point base: critical -20, high -10, warning -5, info -1.
3. The letter grade is gated by max severity AND total score: any critical → F; any high → at most D; any warning → at most C; otherwise A/B by score band.
4. Per-category sub-scores apply the same deduction formula to that category's findings only — so you can see WHICH risk surface drove the loss.

Rule matches are augmented by an LLM-based semantic pass when an LLM endpoint is configured. The semantic pass uses rule IDs SEM-001 … SEM-008.

When an LLM endpoint is configured the skill is also probed with a 15-attack adversarial corpus (5 classes × 3 prompts), each judged by a separate LLM call. Failed classes surface as rule IDs AR-001 … AR-005.

Engine + rule set provenance:

• Engine version: 0.2.0
• Rule set version: 1.1.0
• Commit: unknown
• Domain config: general
• Audited at: 2026-07-11T20:45:55.930068Z
• Rules applied: 36 static rules (full registry below)

Known limitations of this report

• False positives are possible. A SKILL.md documenting a dangerous pattern (e.g. an audit skill explaining curl | sh) will match the rule even though the skill's intent is to detect, not execute. Read the matched lines before reacting.
• False negatives are guaranteed in narrow ways. Patterns obfuscated by string concatenation, environment variable indirection, or non-English equivalents will slip past regex.
• Baseline sample size. Same-skill trend analysis (§ Historical baseline) gets meaningful with n≥3 prior audits. With fewer priors the stddev band is widened to avoid false out-of-band signals.

About TAR Engine

TAR Engine is an OSS "wish machine" with built-in audit. Speak a goal; the engine plans, runs and audits skills inside its own container. BYOK. — github.com/qingxuantang/tar-engine