640 commits

This branch

This branch

All branches

Author	SHA1	Message	Date
clawdie	7f0f06d36a	refactor: rename golden tests → fixtures (consistent positive framing) (#267)	2026-06-28 10:22:17 +02:00
clawdie	5b3511ae86	refactor: rename zot_rpc_smoke → zot_rpc_driver (#266)	2026-06-28 10:01:23 +02:00
clawdie	743e349a13	fix(test): zot_rpc_smoke — remove #[ignore], fall back to PATH (#265)	2026-06-28 09:45:14 +02:00
Sam & Claude	274652a9fb	feat: T2.x eval harness + RPC task dispatch (#264) ... Co-authored-by: Sam & Claude <hello@clawdie.si> Co-committed-by: Sam & Claude <hello@clawdie.si>	2026-06-28 08:43:28 +02:00
clawdie	4fc02a6b28	refactor: rename clawdie-hw-probe → clawdie-system-probe (#263)	2026-06-28 08:34:22 +02:00
clawdie	3b826e1ab2	feat(schema): add blender_available detection to derive_capabilities() trigger (#262)	2026-06-28 08:31:47 +02:00
clawdie	8de42704ae	feat(daemon): add blender detection to probe_capabilities() fallback (#261)	2026-06-28 08:17:55 +02:00
clawdie	0a1c6b3604	feat(daemon): direct ollama/llama.cpp probing fallback in probe_capabilities() (#260)	2026-06-28 08:13:00 +02:00
clawdie	47d43e8d46	Merge pull request 'fix: cargo fmt drift on probe_capabilities call' (#259) from fix/socket-fmt-drift into main ... Reviewed-on: #259	2026-06-28 07:38:14 +02:00
Sam & Claude	a47282d908	fix: cargo fmt on probe_capabilities call ... PR #258 introduced a long line that cargo fmt wants reformatted. No functional change — just formatting. Sam & Claude	2026-06-28 07:13:54 +02:00
clawdie	8071bd0d0d	Merge pull request 'feat(daemon): dynamic agent capabilities via clawdie-hw-probe --capabilities' (#258) from feat/dynamic-agent-capabilities into main ... Reviewed-on: #258	2026-06-28 01:41:14 +02:00
Sam & Claude	3b25668c9e	feat(daemon): dynamic agent capabilities via clawdie-hw-probe --capabilities ... Replaces hardcoded ["shell","freebsd","code"] with a call to clawdie-hw-probe --capabilities during autospawn agent registration. probe_capabilities(): - Runs clawdie-hw-probe --capabilities (cross-platform, < 100ms) - Parses the JSON array from stdout: ["shell","freebsd","code"] - Falls back to ["shell","code"] if probe unavailable Platform examples: FreeBSD: ["shell","freebsd","code"] Linux: ["shell","linux","code","vulkan"] This is Phase 1b of the dynamic capabilities plan. The hw-probe change (Phase 1a) is in clawdie-iso feat/hw-probe-capabilities. Phase 2 (T2.x) adds ollama/llama.cpp/cuda detection to the same probe pipeline — no new env vars, no protocol changes.	2026-06-28 01:38:52 +02:00
clawdie	1681eeca60	Merge pull request 'fix(astro): EN index reads from src/content, not ../../docs/wiki' (#257) from fix/astro-index-path into main ... Reviewed-on: #257	2026-06-28 01:34:01 +02:00
clawdie	2f0f9a2d24	fix(daemon): extract hardcoded values in autospawn registration (#254)	2026-06-28 01:30:33 +02:00
clawdie	56af070b6e	Merge pull request 'fix(daemon): register autospawned agent in store for task routing' (#253) from fix/autospawn-agent-registration into main ... Reviewed-on: #253	2026-06-28 01:07:02 +02:00
Sam & Claude	8df6613410	fix(astro): EN index reads from src/content, not ../../docs/wiki ... Bug: EN index.astro used path.resolve('../../docs/wiki') which only works in the dev environment. Production builds from staged copy at astro/wiki/src/content/ — the other 3 pages already used this. Added 'sl' to EXCLUDE so Slovenian pages don't leak into EN index. Verified: 28 EN pages, zero SL leaks. model-selection-and-eval.md and hive-routing.md render correctly at live wiki.clawdie.si.	2026-06-28 00:59:13 +02:00
Sam & Claude	4f3f9f1a4a	fix(daemon): register autospawned agent in store for task routing ... After autospawn spawns zot, the agent was never registered in the local SQLite store. The scheduler only routes tasks to registered agents — so every task stayed stuck in 'queued' forever. This commit adds a store.register_agent() call right after the spawn succeeds, before the bootstrap RPC prompt. The agent is registered with [shell, freebsd, code] capabilities and the current hostname. Uses try_lock() to avoid blocking the daemon on store contention.	2026-06-28 00:58:31 +02:00
clawdie	5d8016eaf0	Merge pull request 'style: fix pre-existing Prettier drift in 3 docs (structured-reports, cost-dashboard, mother-hive)' (#251) from fix/prettier-drift-3-files into main ... Reviewed-on: #251	2026-06-28 00:25:13 +02:00
clawdie	9ac7d39d30	Merge pull request 'fix/skills-pf-validate-cleanup' (#250) from fix/skills-pf-validate-cleanup into main ... Reviewed-on: #250	2026-06-28 00:23:52 +02:00
Sam & Claude	40f091135d	fix(skills): remove duplicate PF validate line in freebsd-admin SKILL ... The PR added a 'validate PF before reload' bullet in the Controlplane service ports subsection, but the original file already had one at the end using the FreeBSD-native 'service pf reload'. Keep only the one at the bottom — avoids confusing operators with two different reload commands. Sam & Claude	2026-06-28 00:20:33 +02:00
Sam & Claude	43c43a4848	feat(skills): fail2ban-tailscale + freebsd-admin PF rate limiting ... fail2ban-tailscale (new skill): Root cause: key negotiation triggers password-fallback, fail2ban bans IP Path A: PasswordAuthentication no — one line, zero maintenance Path B: Specific fleet IP whitelist — if passwords must stay on Path C: Both — production hardening Security: do NOT whitelist 100.64.0.0/10 (trusts every tailnet) FreeBSD PF equivalent: max-src-conn-rate + overload table Platform table: Linux fail2ban / FreeBSD PF / Mother PF freebsd-admin (PF SSH rate limiting): max-src-conn-rate 5/60 + overload <ssh_brutes> table Manual operations: show, delete specific IP, flush Cross-reference to fail2ban-tailscale skill Rule placement guidance (block drop all last, pass out first) Wiki-lint: 187 refs, 0 failures. Prettier 3.8.4: clean.	2026-06-28 00:15:44 +02:00
clawdie	2406441b4d	fix: remove legacy references — Rust source + agent skills (5 files) (#249)	2026-06-28 00:10:50 +02:00
clawdie	fd1fc6352d	docs: remove legacy references — positive framing pass (11 files) (#248)	2026-06-28 00:07:17 +02:00
clawdie	7d7bc5ca33	style: fix pre-existing Prettier drift in 3 docs (structured-reports, cost-dashboard, mother-hive) (#247)	2026-06-28 00:05:12 +02:00
clawdie	39f74b1c8c	Merge pull request 'docs(hive-routing): fleet SSH reliability — password off, agent keys, PF limits' (#246) from fix/ssh-agent-keys-persistence into main ... Reviewed-on: #246	2026-06-27 23:51:01 +02:00
123kupola	fece199390	style: fix pre-existing Prettier drift in 3 docs (structured-reports, cost-dashboard, mother-hive)	2026-06-27 23:47:50 +02:00
123kupola	780b06de56	docs(hive-routing): fleet SSH reliability — password off, agent keys, PF limits ... Covers the three root causes discovered during osa→debby SSH setup: 1. PasswordAuthentication no — removes fail2ban's reason to exist 2. AddKeysToAgent yes — prevents ksshaskpass popups on reconnect 3. FreeBSD PF rate limiting — defense in depth for osa Sam & Hermes	2026-06-27 23:22:23 +02:00
clawdie	20408bf7f3	Merge pull request 'fix(dashboard): restore dual-proof lightbox — screenshot + text badges' (#245) from fix/dual-proof-lightbox-restore into main ... Reviewed-on: #245	2026-06-27 22:36:37 +02:00
Sam & Claude	ff9c8511f9	fix(dashboard): restore dual-proof lightbox — screenshot + text badges ... Restores structure removed by PR #242: <img id=lb-img> for screenshot proofs (future — daemon doesn't populate yet) <pre id=lb-text> for proof_text (glasspane state JSON) Typed badges: ▸ screenshot when screenshot_uuid populated, ▸ text otherwise Co-authored-by: Hermes <hermes@clawdie.si>	2026-06-27 22:34:40 +02:00
clawdie	80543c5f46	Merge pull request 'docs/model-selection-and-eval' (#244) from docs/model-selection-and-eval into main ... Reviewed-on: #244	2026-06-27 22:24:25 +02:00
Sam & Claude	b096168aee	docs(wiki): model selection + evaluation harness design ... New wiki page: model-selection-and-eval.md (445 lines) Completes the T2.x trifecta design: - Evaluation harness: 3 modes (self-report, local LLM, cloud LLM) - Model selection: weighted scoring (success rate, cost, capability, latency) - Integration with hive-routing: data flow + implementation phases - 4 implementation phases, ~10 days total, ~570 lines Indexed in both en/index.md and sl/index.md. Follows PR #241 (conflict marker fix) and the now-merged screenshot pipeline. The eval harness provides the feedback loop that makes model-selection decisions data-driven rather than heuristic. Sam & Claude	2026-06-27 22:18:18 +02:00
Sam & Claude	08cdae1c47	fix(tests): cargo fmt on cost_pipeline.rs — PR #243 followup ... Cargo fmt drift in the new cost pipeline integration tests: - Multi-line .args() calls (8+ args per line) - Multi-line assert!() with format strings - Braced if-let-else blocks Sam & Claude	2026-06-27 22:18:18 +02:00
clawdie	6e40944389	Merge pull request 'test: cost pipeline — SSH roundtrip + MCP cost dispatch + typo fix' (#243) from test/cost-pipeline-tests into main ... Reviewed-on: #243	2026-06-27 22:12:08 +02:00
clawdie	ed252ab159	Merge pull request 'fix/dashboard-dead-screenshot-code' (#242) from fix/dashboard-dead-screenshot-code into main ... Reviewed-on: #242	2026-06-27 22:10:45 +02:00
Sam & Claude	0323a9817e	fix(dashboard): remove dead screenshot_uuid badge/code — daemon never populates it ... The screenshot.rs module was never committed to main (PR #239 commit message was aspirational). The daemon pushes only proof_text (glasspane JSON). Removed: - hasScreenshot / ▸ screenshot badge branch - <img> lightbox path (only proof_text <pre> remains) - Orphaned else-if block from dual-badge logic Kept: - screenshot_uuid column in schema (future-proof, harmless) - Comment explaining why only proof_text is checked - JSON panel includes screenshot_uuid (raw field, matches schema)	2026-06-27 22:08:31 +02:00
123kupola	d4aa7f9b9f	test: cost pipeline — SSH roundtrip + MCP cost dispatch + typo fix ... SSH integration tests (gated on COLIBRI_MOTHER_HOST): - ssh_mother_report_task_cost_roundtrip: stdin payload → INSERT 0 1 - ssh_mother_tools_lists_daemon_tools: verify tools discovery - ssh_mother_rejects_unknown_commands: validate allowlist MCP dispatch tests (mocked daemon): - tool_list_task_costs_returns_cost_data: cost fields in output - tool_get_task_returns_cost_fields: task lookup with cost Bug fix: - colibrie_deploy_ → colibri_deploy_ (two MCP tool names) Sam & Hermes	2026-06-27 22:06:56 +02:00
Sam & Claude	bcc24002f8	fix(mother): replace pg_read_file with psql variable interpolation ... pg_read_file('/dev/stdin') requires superuser — colibri role had only INSERT. Use psql -v json_input="$(cat)" + :'json_input' for parameterized JSON ingestion. Works with non-superuser PostgreSQL roles.	2026-06-27 21:43:12 +02:00
clawdie	a5c9da3ef6	Merge pull request 'fix: resolve merge conflict markers in mcp-ssh + schema (PR #240 merge artifact)' (#241) from hotfix/merge-conflict-markers into main ... Reviewed-on: #241	2026-06-27 21:32:02 +02:00
Sam & Claude	3becc6c190	fix: resolve merge conflict markers in mcp-ssh + schema (PR #240 merge artifact) ... Two files had unresolved <<<<<<< HEAD / >>>>>>> markers committed to main: - colibri-mcp-ssh: resolved to JSON proof_text example + screenshot_uuid - mother_schema.sql: resolved to both columns (proof_text + screenshot_uuid)	2026-06-27 21:30:08 +02:00
clawdie	1034842145	Merge pull request 'fix/polish-proof-text-refs' (#240) from fix/polish-proof-text-refs into main ... Reviewed-on: #240	2026-06-27 21:26:00 +02:00
123kupola	d36407766e	docs: replace stale tmux-screenshot refs in cost-dashboard.md	2026-06-27 21:22:47 +02:00
Sam & Claude	b9b7626efe	chore: polish proof_text — stale comments, duplicate column, schema docs, fmt ... - schema: removed duplicate proof_text column, fixed stale tmux-screenshot comment - mcp-ssh: example now shows real proof_text JSON instead of UUID placeholder - daemon: trimmed stale COLIBRI_TASK_SCREENSHOT_UUID comment - wiki/cost-dashboard: updated architecture diagram + JSON example - fmt: trailing whitespace in colibri-mcp lib.rs	2026-06-27 21:20:21 +02:00
clawdie	5702c33c30	Merge pull request 'feat: screenshot proof pipeline — module + proof_text + visual PNG' (#239) from merge/screenshot-proof-pipeline into main ... Reviewed-on: #239	2026-06-27 21:10:39 +02:00
123kupola	339c87ce51	feat: screenshot proof pipeline — module + proof_text + visual PNG ... Combines: - screenshot.rs module (is_enabled, generate_uuid, capture) with CaptureError enum (4 variants: Disabled, SpawnFailed, ExitFailed, NoPng) - proof_text: glasspane text evidence at task exit - screenshot_uuid: visual terminal proof (PNG capture UUID) - Daemon: fire-and-forget capture in heartbeat, both fields in payload - Schema: proof_text + screenshot_uuid columns (no duplicates) - SSH wrapper: INSERTs both fields via NULLIF - Dashboard: dual lightbox — <img> for screenshots, <pre> for text ▸ badge shows type: 'screenshot' or 'text' - Deploy: FreeBSD /etc/crontab (not Linux cron.d) - JSON export: includes both proof_text and screenshot_uuid Sam & Hermes	2026-06-27 21:08:42 +02:00
clawdie	59bd530a81	Merge pull request 'feat/proof-text-inline' (#237) from feat/proof-text-inline into main ... Reviewed-on: #237	2026-06-27 19:22:08 +02:00
Sam & Claude	79b4011c9e	feat: replace screenshot_uuid PNG proof with inline proof_text ... Option C — text proofs instead of screenshots. - daemon: capture proof_text from glasspane at task exit (agent name, state, session, tokens, cost_usd → JSON string) - mother_schema: screenshot_uuid → proof_text TEXT column - colibri-mcp-ssh: INSERT proof_text instead of screenshot_uuid - export-costs.sh: include proof_text in JSON export - dashboard: lightbox shows <pre> text proof instead of <img> openProof() looks up task from DATA by task_id Zero disk overhead — proof_text is inline text (~200 bytes), not a PNG file (~200KB).	2026-06-27 19:09:07 +02:00
Sam & Claude	cb124a014f	feat(deploy): add colibri-deploy crate + MCP tools ... New crate: colibri-deploy - run(target, command) → shell on host or Bastille jail - list_targets() → host + active Bastille jails (graceful sudo fallback) New MCP tools (2): - colibri_deploy_run: run a command on host or jail - colibri_deploy_targets: list available targets Tests: host echo, host fail, list_targets includes host. Tool count: 18 → 20	2026-06-27 18:57:55 +02:00
clawdie	9e8e25d535	Merge pull request 'chore: delete dashboard handoff — Steps 1-3 done, Step 4 filed as task' (#236) from chore/delete-dashboard-handoff into main ... Reviewed-on: #236	2026-06-27 18:19:44 +02:00
Sam & Claude	25f4611772	chore: delete dashboard handoff — Steps 1-3 done, Step 4 filed as task ... Steps 1-2 (schema + daemon plumbing): merged. Step 3 (dashboard HTML): built differently than spec, deployed. Step 4 (OSA setup): filed as create-task on the board — tracked, not rotting in a stale doc.	2026-06-27 18:17:38 +02:00
clawdie	0d4d0f7fb1	Merge pull request 'chore/merge-all-uncommitted' (#235) from chore/merge-all-uncommitted into main ... Reviewed-on: #235	2026-06-27 18:11:48 +02:00