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 alex-jb/shadow-perception-mcp --skill shadow-perception-powerbigit clone --depth 1 https://github.com/alex-jb/shadow-perception-mcpWrote 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/alex-jb/shadow-perception-mcp/shadow-perception-powerbi)<a href="https://agentmods.dev/skills/alex-jb/shadow-perception-mcp/shadow-perception-powerbi"><img src="https://agentmods.dev/badge/skills/alex-jb/shadow-perception-mcp/shadow-perception-powerbi/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/alex-jb/shadow-perception-mcp/shadow-perception-powerbi"><img src="https://agentmods.dev/badge/skills/alex-jb/shadow-perception-mcp/shadow-perception-powerbi.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 | $0.00108 | $0.00804 |
| Opus 5 | $0.00054 | $0.00402 |
| Sonnet 5 | $0.00022 | $0.00161 |
| Haiku 4.5 | $0.00011 | $0.00080 |
Grade A, and why
shadow-perception-powerbi 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 9d 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 — 71 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Shadow Perception — Power BI vertical
Five-voice deterministic pre-publish council for Power BI semantic models. Deterministic — no LLM call in the verdict path.
When to use
The user's request is any variant of:
- "Is this DAX safe to publish?"
- "Any governance issues with this .bim?"
- "Did I introduce a divide-by-zero?"
- "Is USERPRINCIPALNAME() actually inside an RLS role?"
- "Any duplicate measures?"
Or the user points at a .bim / .tmdl / .dax file and asks for feedback before publish to Power BI Service.
What it does
shadow_perception_scan(file_path)— parses .bim (Tabular Model JSON), .tmdl (Tabular Model Definition Language), or .dax (raw DAX). Extracts measures, RLS roles, DAX signals (DQ01 bare division / DQ02 BLANK without COALESCE / GOV01 USERPRINCIPALNAME / PERF01 nested CALCULATE / PERF02 EARLIER), and duplicate measure names.shadow_perception_council(observation)— 5-voice council:- Data Quality — DQ01 / DQ02 (divide-by-zero, null propagation)
- Governance — GOV01 (USERPRINCIPALNAME without RLS role = bypassable)
- Semantics — duplicate measure names, empty model
- Compliance — PII column references (email / phone / SSN / DOB / address / passport)
- Ops — PERF01 (nested CALCULATE ≥ 4) / PERF02 (EARLIER pattern)
shadow_perception_attest(observation, verdict)— Ed25519 signs.
Non-goals
- Not a full DAX linter. Style / naming / SUMX vs SUM out of scope. DAX Studio + Tabular Editor own those.
- Not a query optimizer. Rough performance smells only; use DAX Studio for real timings.
- Not a security auditor. RLS role definitions are advisory-flagged, not enforced.
Companion product
Same Ed25519 attestation primitive as alex-jb/shadow-mentor.
Reference
- Reg BI + GDPR Art. 32 access control by construction
- Microsoft Best Practice Analyzer rules (parity target)
- Tabular Editor: https://tabulareditor.com/
- pbi-tools: https://github.com/pbi-tools/pbi-tools
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.
- 9d ago First seen · 71 lines · 108 tokens per session scan A 5b32ebd46b99
shadow-perception-powerbi is a skill published in the GitHub repository alex-jb/shadow-perception-mcp (0 stars, last pushed 2mo ago), licensed MIT. It adds 108 tokens to every session and 804 once invoked, about $0.0005 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 skills, from other repositories
agent-passport-system
Enforcement and accountability layer for AI agents. Bring your own identity (did:key, did:web, SPIFFE, OAuth, did:aps). Gateway enforcement boundary, monotonic narrowing, cascade revocation, spending controls, data lifecycle, observation governance (telemetry scopes, derivation rights, behavioral memory). Use when…
agent-upgrade
The one skill an AI agent adds to itself to become verifiable, accountable, and interoperable. It gives the agent an Ed25519 identity it controls, provable operation within the authority delegated to it, its own cryptographic receipt chain it can cite, the ability to communicate, delegate, and transact with other…
core-skill
The base layer an AI agent runs on to be verifiable, accountable, and able to work with the world. One slug adds a controlled Ed25519 identity tied to the principal behind the agent, scoped authority that only narrows, a signed receipt for every action the agent can carry across sessions and handoffs, and a connection…
model-citizen
Early release. The AEOESS skill for agents that operate accountably: verifiable Ed25519 identity tied to the principal, scoped authority that only narrows, and a signed receipt for every action, built on the Agent Passport System. The agent acts only within delegated authority. Apache-2.0. This is an early release…
powerbi-expert
Expert-level Power BI, DAX, M language, data modeling, Power Query, report design, and paginated reports. Use when the user mentions DAX, Power Query, BI, Microsoft platforms, analytics, or data modeling, or when the task involves DAX Fundamentals, Advanced DAX, Row-Level Security, or Report Design.
mcp-engine-dax-performance
Use when a DAX query, measure, or visual is slow, when interpreting analyze timings, Storage Engine / Formula Engine splits, or query plans, when VertiPaq storage size or cardinality drives cost, or when the user wants a tuning pass or before/after benchmark. For wrong values, use mcp-engine-dax-debugging; for writing…