Internal Comms Skill
A skill for writing internal communications — status reports, 3P updates, company newsletters, FAQs and incident reports — in your company's preferred formats.
The skill is prompt-only guidance: no code execution, no privileged permissions, no sensitive-data flows of its own, so the risk surface is small. However, the example files direct the agent to pull from Slack, email, Drive and calendar without explicit constraints on data sourcing, privacy or confirmation; rollback/isolation concepts are neither applied nor declared, so full marks are not awarded.
Instructions are self-consistent, the four comm types map cleanly to example files, and there is a fallback to ask the user when no guideline matches. But the skill depends entirely on the host agent having Slack/Drive/email tools, offers no structured failure feedback when they are absent, and has no tests or edge-case handling; static ceiling is 10.
Trigger keywords (3P updates, newsletter, FAQs, status reports, etc.) are clear, audience and scenarios are explicit, and a clarification path exists for unmatched types. But capability boundaries (limited value without internal tooling access) are only implied inside examples, and there is no Chinese-language or environment-fit note, hence a mid-high score.
SKILL.md is well structured with progressive disclosure into example files, and license metadata is present. But the referenced LICENSE.txt is not present in the repo (only the repo-level MIT LICENSE), there is no versioning, changelog or maintenance-ownership statement, and the newsletter example contains typos (perhgaps), so governance is incomplete.
Concrete, directly usable templates are provided (strict 3P format, newsletter sections and examples, FAQ Q&A format), and the core writing task is plausibly completed. But much of the value depends on pulling context from external tools, output quality cannot be statically verified, and comparative-benefit evidence is thin; static ceiling is 7.
Content and formatting rules are auditable in the source files, constituting primary material. But there are no skill-specific tests, no third-party execution evidence, no independent corroboration, and the content appears derived from Anthropic's skills repository without attribution; static ceiling is 5.
- The skill directs the agent to read Slack, email, calendar and Google Drive data; users should mind enterprise data privacy and access authorization boundaries.
- The referenced LICENSE.txt is not included, and the content appears ported from Anthropic's official skills without attribution; verify licensing before commercial use.
- Effectiveness depends heavily on the host agent having Slack/Drive-like tools configured; without them the skill mostly polishes user-provided content.
- No Chinese-language adaptation notes, and the embedded company assumptions (e.g., 1000+ employees) reflect an English enterprise context.
- This is a static source review: no tests were executed and output quality is unverified.
What does this skill do, and when should you use it?
This is one of 15 skills bundled in the Mini-Agent repository, located at mini_agent/skills/internal-comms/. It contains no executable scripts; instead it is a set of writing resources: it identifies the communication type in a request, loads the matching format-and-tone guideline from the examples/ directory, and writes the document accordingly. Supported types include 3P updates (Progress/Plans/Problems), company newsletters, FAQ responses, status reports, leadership updates, project updates and incident reports. When a request doesn't match an existing guideline, the skill asks for clarification rather than guessing. The surrounding repository is MIT-licensed, installed via uv, and runs against the MiniMax M2.5 model.
The skill operates purely through instructions: 1) it identifies the communication type from the user's request (e.g., 3P update, newsletter, FAQ); 2) it loads the corresponding guideline file from examples/ — examples/3p-updates.md, examples/company-newsletter.md, examples/faq-answers.md, or the fallback examples/general-comms.md; 3) it follows that file's instructions for formatting, tone, and content gathering to produce the communication document. It reads the repository's guideline files and the user's request, and outputs formatted internal-communication text.
- A team lead who must submit weekly Progress/Plans/Problems updates and wants a consistent reporting format.
- An internal comms specialist drafting company-wide newsletters who needs a repeatable tone and layout.
- An engineering team writing an incident report after an outage, guided by a ready-made structure.
- A project manager publishing project status updates or leadership briefings in the company's usual style.
- Support or ops staff answering FAQs who want consistent answers across the team.
What are this skill's strengths and limitations?
- Covers 7 common internal-communication scenarios, including the standardized 3P update structure.
- Pure instruction-based skill with no script dependencies, giving good cross-platform and cross-model portability.
- Includes a fallback guideline (general-comms.md) and asks for clarification on unmatched types rather than guessing.
- Backed by an MIT-licensed, actively maintained repository with documented install and upgrade flows.
- The actual content of the example guideline files is not shown, so format quality can't be assessed up front.
- Guidelines reflect one specific company's format preferences and may not match other organizations' norms.
- No skill-specific test suite; the repository's tests target the agent core, not the skills' output quality.
- No independent versioning or changelog at the skill level — maintenance info exists only at the repository level.
How do you install this skill?
The skill ships as part of the Mini-Agent repository. Quick install: uv tool install git+https://github.com/MiniMax-AI/Mini-Agent.git, then run the official setup script and fill in your MiniMax API Key and API Base in ~/.mini-agent/config/config.yaml. Development mode: git clone the repository, run uv sync, then git submodule update --init --recursive to initialize the skills. The skill lives at mini_agent/skills/internal-comms/; the source does not document placing the skill folder into other Agent Skills-compatible clients.
How do you use this skill?
In Mini-Agent, simply make an internal-communication request such as "Write this week's 3P update" or "Draft an incident report for the service outage." The skill identifies the type, loads the matching examples/ guideline, and produces the document in that format. If the type doesn't match an existing guideline, it asks you for the desired format or more context.