Compress the skill.
Keep the behaviour.
SkillZip Pro is an evaluation-free way to shrink evolving agent skill bundles while preserving how they load, what they can reach, and which entries can run on their own.
route + contract
A production skill is a folder, not a prompt.
As an agent evolves, useful instructions get copied across roots, references and branches. SkillZip Pro removes the repetition, but keeps each branch reachable and each public entry usable.
Builds a typed map of the bundle, finds safe reuse, and emits an ordinary directory the existing harness can read.
Less shipped content, less repeatedly loaded context, and fewer tokens spent re-reading the same rules.
Every removal needs evidence. Routing, contracts, locked files and the materialised output are audited before publication.
Self-evolving research agents, tool-use skills, content moderation, spreadsheet agents, coding assistants, and any agent whose knowledge is split across files and branches.
Agents do not load the whole bundle at once.
They progressively disclose it: discover the skill, activate the root, then open only the resources needed by the current task.
That means a token saved from an on-demand reference is not equivalent to a token saved from the root. The root is paid for repeatedly.
A big ratio can hide a broken skill.
The paper tests three tempting shortcuts. Each optimises one view of the bundle while damaging another.
Short root, dead branches
Rewriting only SKILL.md can remove the routing table. Files remain on disk, but the agent can no longer discover them.
Huge ratio, lost knowledge
Concatenate all files, compress the text, then map it back. Progressive-loading boundaries disappear, and branch-specific content can be deleted.
Less storage, more runtime
Move shared text into the root and every task pays for it. A rare branch can become part of every activation.
Share at the right scope
Factor repetition into the lowest safe activation scope, move long guarded sections into capsules, and keep route contracts intact.
Two pillars make compression useful.
Activation-aware cross-file compression
Find savings across the resource graph without promoting rare knowledge into the root.
- Remove text already guaranteed by a signed host contract.
- Share exact rules inside the scope that actually needs them.
- Move long, explicitly guarded sections into on-demand capsules.
Cross-file routing preservation
Treat navigation as a contract, not ordinary prose.
- Lock routing instructions and reference-bearing lines.
- Keep relative paths and public entry points discoverable.
- Re-read the emitted directory and reject unsafe candidates.
One compression ratio is not enough.
SkillZip Pro reports catalog, activation, path and deployment cost separately, then optimises a weighted objective.
Moving 500 tokens out of an on-demand file can be good. Moving those same 500 tokens into the always-loaded root can make every task worse.
The paper uses a default storage weight λ = 0.05, but reports all layers even when optimising the combined objective.
Choose when to compress, then choose where the result lives.
The axes are independent. Click a mode to see the paper's intended use.
Where
Smallest shipped bundle and lower steady-state cost.
Canonical bundle stays byte-identical.
Zip-on-Write with periodic global repack.
Cached task-specific views for moving targets.
Evidence, not deletion count, sets the safe ceiling.
Each removal carries one witness. Stronger witnesses allow more savings, but the paper keeps interface contracts atomic and rejects anything it cannot justify.
Literal containment. The removed text survives byte-for-byte at another reachable location.
Coverage witness. A merged or shared representation preserves every protected literal and polarity.
Entailment witness. A frozen checker confirms the decision-relevant content is already expressed elsewhere.
No environment contract, no host-entailment deletion. If an audit fails, the method falls back to a verbatim bundle or raw execution closure.
It is a compiler with a hard stop.
The optimiser can propose aggressive changes. The disk-level audit decides whether anything is allowed to ship.
Publish the ordinary relative-file directory and save the compressor state for the next evolution patch.
Keep the patched source or raw closure. Compression is allowed to lose savings, not to lose a valid update.
SkillZip Pro was the only compressor that cut all four costs and kept quality.
Self-evolution manufactures the redundancy it can later remove.
Across the paper’s evolved libraries, about 55% of each new round repeated text already written. A few universal rules absorbed most of that reuse.
Share of shipped tokens removed from round 2 to round 15.
Reuse captured by the top two of seven shared modules.
The paper reports 10.4% lower pooled, task-paired runtime tokens over 200 audits, with 89% accuracy in the compressed runs.
| Configuration | Deployment | Runtime | Accuracy |
|---|---|---|---|
| Uncompressed | reference | reference | 88-92% |
| Pro + W1/W2 | 13.8% lower | 7.1% lower | 89% |
| Pro + W3 | 32.7% lower | 6.8% lower | 88% |
| Pro + W3 inside root | 38.1% lower | 10.4% lower | 89% |
The unprotected 71.4-75.8% configurations lost 18-26 accuracy points, primarily through additional false positives.
The right lifecycle depends on how the bundle changes.
Persistent compression pays its build cost once and stays cheaper across repeated direct calls.
Transient compression can rebuild only the affected execution view and preserve the canonical bundle.
Continual persistent compression uses local repair plus triggered global repacks to control drift.
Structural, not universal
The guarantee protects explicit contracts and routes. It does not prove every possible model interpretation remains equivalent.
Routing gets more important
Shared modules reduce duplication, but the executor must correctly follow the extra links. Cross-model results were mixed.
Evidence is a constraint
Without a stable, typed environment contract, host-entailment pruning is disabled. Unclear deletions are refused.
Start with the bundle you already have.
Use the stepper as a practical implementation plan. The exact paper describes the compiler; this sequence turns it into an engineering rollout.
Make the audit the product boundary.
If the candidate is smaller but a route, contract, locked byte or patch disappears, it is not a valid output.
Build a read-only analyser that emits the resource graph, entry contracts, candidate witnesses and four cost layers before it is allowed to rewrite anything.
Good compression understands the shape of execution.
SkillZip Pro’s contribution is not simply shorter text. It treats an agent skill as a living, routed resource graph, and makes every saving answerable to where and how the agent will load it.
Remove repeated knowledge once, at the lowest safe scope.
Reachability and entry independence are first-class constraints.
Deployment savings are useful only when execution cost and quality hold.
What this deck preserves from the paper.
This is an explanatory visual summary, not a reproduction of the formal proofs, all tables, or the full appendix.
Reported figures are taken from the supplied PDF, especially Tables II-III, IV-VI, IX-XIII and XXIII.
Use cases and implementation framing are explanatory extensions. They are not additional empirical claims from the paper.
Prepared from the supplied file: “SkillZip Pro-Execution-AwareDynamicCompression of ProgressivelyLoadedSkills for Self-EvolvingAgents.pdf”.