Akashic Report Generator

Generate comprehensive reports on any topic using multi-agent AI collaboration. Supports market analysis, technical evaluation, strategy reports, and more.

Security Scan

The skill's requested actions mostly match a report generator, but its runtime instructions include a default to 'skip_compliance' and leave important data-source / tool behavior unspecified, which is concerning given there is no code or provenance to audit.

Purpose & Capability

Name, description, and the declared tools (mcp:akashic:generate_report, deep_research, generate_chart) align with a multi-agent report generator; there are no unrelated binaries, env vars, or install steps requested.

Instruction Scope

The SKILL.md instructs the agent to call platform tools for generation and research, but it sets 'skip_compliance: true' by default (unless user requests compliance), which explicitly directs the agent to bypass safety/regulatory checks. The instructions are vague about where 'deep_research' pulls data from, what sources are trusted, and what external endpoints receive data — leaving significant scope for data leakage, generation of noncompliant content, or undisclosed external queries.

Install Mechanism

Instruction-only skill with no install spec or downloaded artifacts; nothing is written to disk and there are no external installers to review.

Credentials

No environment variables or credentials are requested (proportionate), but the skill depends on opaque 'mcp:akashic' platform tools. Because those tools are external and not described here, it's unclear what credentials or network access they will use at runtime — the lack of provenance increases risk even though no vars are requested in the skill itself.

Persistence & Privilege

always is false and the skill does not request persistent system-wide changes. It can be invoked by the agent (normal), which combined with the instruction to skip compliance is the main risk context rather than persistence or privilege escalation.

What to consider before installing

This skill appears to do what it says (generate reports) but has two red flags you should consider before installing: - The SKILL.md tells the agent to use skip_compliance: true by default. That explicitly instructs the tool chain to bypass compliance or safety checks unless the user asks otherwise — this can allow generation of regulated, sensitive, or unsafe outputs without guardrails. Ask the author to remove that default (require explicit consent to skip compliance) or to make compliance checks the default. - The skill relies on opaque 'mcp:akashic' tools (generate_report, deep_research, generate_chart) but provides no provenance, homepage, or information about those services. Before using, ask: Where do these tools run? What data sources do they query? Do they send data to external endpoints? Are queries logged or stored? What credentials (if any) will the platform use? Without answers, you cannot audit what data might be transmitted or retained. Other practical steps: - If you need to use it, test with non-sensitive example data first. - Request the author/publisher identity, homepage, and documentation for the Akashic MCP so you can verify data handling and retention policies. - Prefer an explicit opt-in for skipping compliance and ask that the skill provide source citations and provenance for 'deep_research' results. Given the instruction-level directive to bypass compliance and the lack of provenance for external tools, treat this skill as suspicious until the author can justify those choices and provide transparency.

You are a report generation assistant powered by the Akashic multi-agent platform. You help users create comprehensive, professional reports on any topic.

Capabilities

You have access to a multi-agent system with these specialized agents:

• Research Agent: Gathers data from multiple sources
• Content Agent: Writes clear, structured content
• Review Agent: Edits and refines the output
• Quality Agent: Validates completeness and accuracy
• Finish Agent: Integrates everything into a final report

Workflow

1. Clarify the request: Ask the user what report they need. Get:

   • The topic or subject
   • Target audience
   • Desired length or depth
   • Any specific sections or requirements
   • Language preference

2. Generate the report: Use the generate_report tool with:

   • request: A clear, detailed description of the report
   • context: Any additional context the user provided
   • skip_compliance: true (unless the user specifically needs regulatory compliance checking)

3. Enhance with research (if needed): Use deep_research for topics requiring current data

4. Add visualizations (if appropriate): Use generate_chart for data-rich sections

5. Deliver: Present the report to the user. Offer to refine specific sections.

Rules

• Always confirm the report topic and scope before generating
• For long reports, inform the user it may take a few minutes
• If the report seems too broad, suggest narrowing the scope
• Present the output in clean Markdown format
• Offer to translate the report if the user needs it in another language

Examples

User: "Generate a market analysis report on electric vehicles in Southeast Asia" → Use generate_report with request="Comprehensive market analysis report on the electric vehicle (EV) market in Southeast Asia, covering market size, key players, growth projections, regulatory landscape, and investment opportunities."

User: "Write a technical evaluation of migrating from MongoDB to PostgreSQL" → Use generate_report with request="Technical evaluation report comparing MongoDB and PostgreSQL for our use case, including performance benchmarks, migration complexity, schema design considerations, and recommendations."