Elaborate forme/prose/memory skill docs with fuller guidance, and seed
rfcs/ (scheduled execution, feedback loops, env declaration, composite
instantiation) plus a std/ README stub.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Propose making `run` a value type in requires/ensures contracts so the VM
can track provenance edges between runs, reconstruct DAGs, detect staleness,
and support fan-in composition (run[]).
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
New file:
- deps.md — authoritative spec for dependency resolution via GitHub repos
Updated files:
- SKILL.md — prose install command routing, .deps/ and prose.lock in file locations
- forme.md — .deps/ as resolution step 3 with path mapping examples
- state/filesystem.md — .deps/ and prose.lock in directory structure
- help.md — install commands, quick start, FAQ
- guidance/patterns.md — dependency-management and std-shorthand patterns
- guidance/antipatterns.md — unpinned-dependencies and vendored-copies
Key design decisions:
- GitHub is the registry (no custom server)
- prose.lock is plaintext (owner/repo sha), committed to git
- .deps/ is gitignored, reproducible from prose.lock
- std/ is shorthand for openprose/std/
- Transitive deps via multi-pass with cycle detection
- Version conflicts auto-resolve to newer SHA with warning
- Explicit prose install required (no auto-install on prose run)
Standard library programs migrated to github.com/openprose/std
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Why: OpenProse maps well onto Codex subagents, but Codex users need explicit docs and example files for recursive execution, alias resolution, and audit-role fanout without relying on machine-local configuration.
All 9 lib/*.md programs now have full v2 contracts:
- requires/ensures with detailed descriptions
- errors sections for failure modes
- strategies for non-obvious execution guidance
- invariants for correctness constraints
- inline ## service definitions for multi-service programs
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
List all migrated v2 examples organized by category, note the
v1-archive, and add counts: 35 migrated, 15 archived, 5 new.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Migrate all 9 library programs: inspector, calibrator, cost-analyzer,
error-forensics, profiler, program-improver, vm-improver, project-
memory, and user-memory. Each becomes a .md file with contracts,
services lists, and appropriate persistence declarations.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Complex execution-block programs: captain's chair full (29), feature
factory (35), The Forge browser build (37), architect by simulation
(39), and language self-improvement (47). All use Level 3 execution
blocks with explicit let/call control flow.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Index run scaffolded new READMEs for all subdirectories. Review run
added the missing profiler.prose entry to skills/open-prose/lib/README.md
(cost, token usage, and time profiling for completed runs).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add ◇ attention markers for Anthropic models throughout SKILL.md
- Add install instructions for remote readers at top
- Expand "You are the Prose Complete system" with explicit capability framing
- Add Post-Execution section with guidance for success/failure cases
- Create SOUL.md template for agents to persist OpenProse awareness
- Add SOUL.md to file locations table
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
After fetching a page, websh now automatically prefetches linked pages
in the background. This makes `follow` and navigation feel instant
since content is already cached when needed.
- Add state/crawl.md with full crawl agent design
- Add prefetch, crawl, queue commands
- Configure via EAGER_CRAWL, CRAWL_DEPTH, CRAWL_SAME_DOMAIN env vars
- Priority: earlier links, same-domain, main content first
- Respects rate limits and robots.txt
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- HN is first in all suggestion lists
- Tab completion should suggest `cd https://news.ycombinator.com` first
- Added note: HN is the "hello world" of websh
- Updated example session to reflect async flow
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Core principle: main thread never blocks.
All heavy operations delegate to background haiku subagents:
- Network fetches
- HTML parsing / extraction
- Content processing
- Multi-page operations (find, crawl, tar)
- Initialization
- Monitoring (watch)
User always gets prompt back instantly. Commands gracefully
degrade if background work isn't complete.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
cd is now fully async:
- User gets prompt back immediately with "(fetching...)"
- Background haiku task does fetch + extract
- Never blocks the main thread
- Commands gracefully degrade if content not ready
Curated bookmarks (removed mainstream news):
- hn, lobsters, tildes, wiby, marginalia, wiki, sourcehut, are.na
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- No more bare "error: no page loaded" - suggest sites to visit
- Pre-populated bookmarks: hn, wiki, lobsters, tildes, npr, cnn, wiby
- Pre-populated aliases for quick navigation
- Helpful hints in empty states instead of "(none)"
- All suggested sites are public/accessible (no walled gardens)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Spawn background haiku task for workspace init.
User sees banner + prompt immediately, can start typing.
Commands work gracefully with empty defaults until init completes.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
A Unix-like shell for navigating and querying the web:
- URLs as paths, DOM as filesystem
- 80+ commands: ls, cat, grep, find, mount, watch, etc.
- Async haiku extraction for rich cached content
- Job control, environment, mounting APIs
- Natural language support - infers intent for unknown commands
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Telemetry was collecting anonymous usage analytics but is no longer needed.
Removes all telemetry-related code, config, and documentation including
USER_ID/SESSION_ID tracking, analytics endpoint calls, and migration steps
for legacy telemetry state files.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Adds two new programs to the standard library that complete the
recursive improvement loop:
- vm-improver.prose: Analyzes inspection reports and proposes PRs
to improve the OpenProse VM specification and implementation
- program-improver.prose: Analyzes inspection reports and proposes
PRs to improve .prose source code (either in lib/ or external)
Together with inspector.prose, these enable a feedback cycle where
execution data drives improvements to both the VM and programs.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add progressive disclosure: quick triage before deep scan
- Model tiering: Sonnet for checklist work, Opus for hard analysis
- Parallel scanners with graceful degradation (on-fail: continue)
- Persistent historian agent for scan history across runs
- Customizable inputs: mode (quick/standard/deep), focus, skill_filter
- Reusable blocks for scan workflow components
- Structured JSON output contracts for all agents
- Early alerting for critical findings mid-scan
- Batched processing to respect parallelism limits
irl-dan