what is sunlc_skills fr?

sunlc_skills is a personally maintained collection of Claude Code skills built around two scenarios that come up when working with an existing codebase: taking over development on a project, and learning to understand a project you did not write. The author describes the skill files themselves as pure methodology, generic across any tech stack, with the actual work carried out by the main session or by dispatched subagents. The collection is split into two complementary toolsets. project-mastery is aimed at someone who needs to write or change code in an unfamiliar project. It scans the codebase, then runs parallel analysis steps covering tech stack and architecture, coding conventions inferred from the actual code, an API index, and build and deployment documentation. It finishes by building an index of that knowledge base and running a lightweight verification pass that samples a few claims per document, with a full opt-in verification mode available for higher stakes projects. learn-project is aimed at someone who wants to understand a project's design and core concepts rather than modify it. It scans for learnable features, lets a human pick which ones to dig into, generates a teaching prompt and document for each chosen feature, and then builds an index of learning paths. The two toolsets write their output to separate folders so they never overwrite each other, and learn-project reuses project-mastery's initial scan instead of repeating it. Installing it means cloning the repository and running a bundled shell script that symlinks each skill folder into the Claude Code or Codex skills directory, since the README says the same skill files work unmodified in both tools. Once linked, a skill can be triggered either by describing the goal in plain language, such as asking to build a knowledge base for a project you are taking over, or by invoking it directly as a slash command in Claude Code. The project also ships two small Python scripts with no external dependencies: one validates the generated knowledge base documents against a schema, and another manages an AI generated versus human edited split inside those documents so automated updates do not overwrite a person's manual notes. The README states the design principles as evidence driven output with no invented claims, and skill files kept free of any project specific names or paths.