The skill dictates agent workflow, forces file writes, generates
Claims to do
Writing Plans: Write comprehensive implementation plans assuming the engineer has zero context for our codebase and questionable taste. Document everything they need to know: which files to touch for each task, code, testing, docs they might need to check, how to test it. Give them the whole plan as bite-sized tasks. DRY. YAGNI. TDD. Frequent commits.
Actually does
This skill processes an input specification to generate a comprehensive markdown implementation plan for software features. The plan details exact file paths, provides code snippets, specifies `pytest` commands for testing, and `git` commands for committing, and is saved to a markdown file (e.g., `docs/superpowers/plans/`). It then suggests using other skills for plan execution.
The skill explicitly dictates the use of other specific 'superpowers' skills and the spawning of 'fresh subagents per task', representing a high degree of control over the agentic workflow and other agents' actions.
REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended) or superpowers:executing-plans ... I dispatch a fresh subagent per task
The skill instructs the agent to save generated plans to a specific file path (`docs/superpowers/plans/YYYY-MM-DD-<feature-name>.md`), which is a direct instruction for file system write operations and could be used for persistence or RAG poisoning.
Save plans to: `docs/superpowers/plans/YYYY-MM-DD-<feature-name>.md`
The skill generates specific shell commands (`git add`, `git commit`, `pytest`) as part of the implementation plan, which are intended to be executed by a downstream agent and pose a risk if the input specification leads to malicious command generation.
Run: `pytest tests/path/test.py::test_name -v` ... `git commit -m "feat: add specific feature"`
The skill uses authoritative language like 'Every plan MUST start with this header' and 'REQUIRED SUB-SKILL', which can semantically manipulate the executing agent's reasoning and reduce its autonomy by enforcing strict adherence to predefined patterns and skill usage.
Every plan MUST start with this header ... REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development
[](https://mondoo.com/ai-agent-security/skills/github/obra/superpowers/writing-plans)<a href="https://mondoo.com/ai-agent-security/skills/github/obra/superpowers/writing-plans"><img src="https://mondoo.com/ai-agent-security/api/badge/github/obra/superpowers/writing-plans.svg" alt="Mondoo Skill Check" /></a>https://mondoo.com/ai-agent-security/api/badge/github/obra/superpowers/writing-plans.svgSkills can read files, run commands, and access credentials. Mondoo helps organizations manage the security risks of AI agent skills across their entire fleet.