Help software preserve its knowledge while its implementation evolves.
Installable skills that teach a coding agent to work so the next change to a codebase
is cheaper and safer than the last one. Built from the Regenerative Software playbook
3.0.0, which ships inside the skill.
Coding agents change code far faster than the knowledge about that code survives. Each session rediscovers the product, guesses at invariants, edits whatever is in front of it, and leaves. Over time the implementation is the only record of what the software promises, the tests no longer prove anything independent of the code, and nobody can say which parts are safe to replace. Every subsequent change becomes more expensive, and a rewrite becomes a gamble.
Regenerative software turns that around. The agent keeps intent (outcomes, invariants, reasons) outside the implementation, judges behavior with evaluations that do not depend on the code they test, leaves provenance (what was run, on which inputs, with which result), and treats replacement of an implementation as a routine operation at a justified boundary rather than a heroic rewrite. Readable, idiomatic source stays; knowledge stops disappearing.
The practical test for any codebase: could another engineer or agent reconstruct an acceptable implementation from the surviving knowledge, evaluate it independently, and introduce it without surprising its consumers? These skills exist to earn that answer boundary by boundary.
1. Install the skill into the project you want to improve. One command, run inside that project; it works for Claude Code, Codex, and other Agent Skills loaders:
cd /path/to/your-project npx skills add https://reg.deploy.qwellco.de --skill regenerative-software
The CLI verifies the archive digest, writes the skill into the agent's skills directory,
and records the digest in skills-lock.json. Commit that file so teammates get
the same copy. Add --agent claude-code or --agent codex to choose
the target. List every skill this host publishes, or install one focused topic:
npx skills add https://reg.deploy.qwellco.de --list npx skills add https://reg.deploy.qwellco.de --skill regenerative-architecture
2. Start a coding session there and state the objective. The entry skill routes to the right topic, discovers the product's journeys and current state, works in bounded slices, verifies with the project's own commands, and leaves a replayable checkpoint. If it does not trigger on its own, name it: "Use the regenerative-software skill to …". For an existing repository, start with:
Use regenerative-software to improve this repository. Discover the important journeys and current working state, choose a valuable bounded slice, preserve supported behavior, implement a justified improvement, and verify it. Embed the seven-primitive operating loop in our actual agent entry points and leave a replayable checkpoint for the next session.
3. Continue from the checkpoint next time. Later sessions reconcile the checkpoint, contracts, and evidence instead of rediscovering the product. After this host is redeployed with a newer skill, refresh the installed copy from the same project:
npx skills update
Modes: translate (default: implement and verify an improvement),
assess (report only), regenerate <unit>,
rehearse <unit>, and compact <scope>. The skills contain
no executable code and require no particular model, MCP server, or framework.
Internal Agent Skills host. Skills version 0.1.0,
playbook 3.0.0, source revision unknown.
Discovery index: /.well-known/skills/index.json.
Build record: discovery-manifest.json.
| Skill | Purpose | Archive |
|---|---|---|
regenerative-software | Design and evolve software so agents can preserve its intent, independently evaluate its behavior, and replace implementations at justified boundaries. Use for regenerative software adoption, agent-maintainable architecture for new projects, repository-wide improvement campaigns, and continued evolution of already adopted repositories. Covers existing codebases, greenfield builds, contracts, state, evaluations, provenance, reconstruction, and retirement. For a routine isolated edit with no regenerative or architectural objective, use the repository's normal workflow. | zip · 80 files · 5ebed09d6347… |
regenerative-architecture | Design new projects and new capabilities around durable intent, coherent ownership, independently changeable boundaries, and behavioral contracts. Use for greenfield regenerative architecture, agent-maintainable project foundations, contract design, or repairing an entangled seam in an existing codebase. Produce a working vertical slice when implementation is requested, with repository guidance and executable knowledge for future agents. | zip · 10 files · 23c480534660… |
regenerative-adoption | Adopt regenerative software practices in existing repositories and embed them in ongoing agent work. Use to discover product behavior, preserve legacy knowledge, prioritize a repository-wide campaign, install concise repository operating instructions, or resume a previous regenerative engagement without losing evidence or overwriting other work. | zip · 13 files · a3a35338ce50… |
regenerative-evaluations | Build meaningful behavioral evaluations that survive implementation changes. Use for regenerative acceptance gates, independent oracles, contract and consumer tests, async lifecycle checks, evaluation failure diagnosis, or runtime model and agent quality assessment. Separate product correctness from test machinery validity and probabilistic quality from hard constraints. | zip · 9 files · 3a53ae23e886… |
regenerative-state | Preserve durable state and external effects through software evolution. Use for regenerative migration design, historical reader/writer compatibility, backup and recovery rehearsal, asynchronous ownership, retries, cancellation, late results, or stateful implementation replacement. Establish what survives code changes and verify it through real consumers. | zip · 7 files · 1a4f7e2bd8c3… |
regenerative-provenance | Make regenerative software changes traceable and replayable with proportionate input identities, execution receipts, canonical documentation, and operational evidence. Use for multi-stage verification, reliable handoffs, generated-document ownership, evidence correction, or distinguishing local checks from release and runtime claims. | zip · 9 files · 64a67688f676… |
regenerative-reconstruction | Exercise implementation replaceability through isolated substitution or fresh-context reconstruction. Use for regenerative survival packs, deletion rehearsals, independent replacements, or claims that another agent can rebuild a unit from retained intent, contracts, state rules, and evaluations. Distinguish source-assisted replacement from pack-only reconstruction. | zip · 10 files · 8f371a76895a… |
regenerative-evolution | Keep regenerative repositories improving across agent sessions. Use for ongoing change pipelines, incident harvesting, selective revalidation, repository-wide continuation, compatibility retirement, deletion, and compaction. Preserve valid fixes and durable knowledge while removing obsolete complexity and testing whether future changes actually become easier. | zip · 11 files · 2139615eb517… |
Download the root archive, extract it, and ask your coding agent to read its
SKILL.md and the files it references. A single Markdown file does not
carry its bundled inputs.