synlynk Changelog

All notable changes to synlynk are documented here.

Format follows Keep a Changelog. Versioning follows Semantic Versioning.

---

Unreleased

Added

• synlynk agent add <name> to retroactively onboard a newly available CLI agent on $PATH, generating its directive file, writing its role fence, wiring workgroup_agents / agent_slots / roles, and seeding a capability baseline. Fixes #277.
• Lazy daily housekeeping on the first synlynk exec of a new calendar day, detecting new PATH agents, re-probing onboarded agents for drift, repairing directive fences, and recording last_housekeeping_date. Fixes #277.

Fixed

• synlynk migrate false-positive MigrationImportError on idempotent re-runs (#276): the loud-fail check treated INSERT OR IGNORE with rowcount == 0 as a hard failure whenever a non-empty source landed zero new rows. On a project that had already been migrated, natural-key collisions (especially roadmap_arcs.version UNIQUE) correctly no-op the insert, so re-import / --recover raised MigrationImportError: 0 rows inserted … roadmap_arcs (N parsed/attempted, 0 inserted) even when every expected row was already present. _migrate_import() now checks natural keys before insert (memory_entries.section, roadmap_arcs.version, roadmap_phases(arc_version, phase_title), cost fingerprint, devlog author+date+title, todo gh_issue already set) and only fails loud when zero rows were inserted and not all attempted rows were already present. Genuine 0-of-N write failures still raise with the same error message prefix.
• Dispatch auto-PR branch detection (#280): _finalize_completed_worktree_job() re-derives the worktree's actual current branch via git branch --show-current instead of trusting the pre-recorded dispatch/<agent>/<job_id> name. When an agent commits on a custom branch (the encouraged human-readable pattern), push and gh pr create --head now target that branch; the job record is updated so synlynk jobs reports the real name. Detached HEAD falls back to the recorded branch without erroring.
• synlynk probe now clears HARNESS_VERSION_DRIFT alerts for the specific agent it just re-probed, so the warning disappears once the recorded installed_version matches the live binary again. Fixes #281.

Release pitch: the operational backbone gets provably reliable and provably accounted for — dispatched jobs finish their own git steps, a 5th zero-cost local agent joins the fleet, story routing gets real capability+quota+cost scoring with a fleet batch scheduler, and every dollar synlynk reports is now either structurally sourced or visibly flagged as an estimate.

Added

Measurement Ledger Hardening — Phase 1 + Phase 2 + display layer (epic #210, PRs #236/#241/#242/#244/#245/#246/#252/#256/#257/#264/#266/#267)

Phase 1 — provenance-tagged cost tracking (PR #236):

• cost_entries gained explicit provenance columns: cost_source TEXT NOT NULL (actual | estimated_token_rate | estimated_tshirt | estimated_manual | legacy_unknown, no default — every INSERT must pass it explicitly) and estimate_basis TEXT. Migration rebuilds the table and backfills historical rows as legacy_unknown.
• _insert_cost_row() (synlynk/db.py) is now the sole sanctioned writer to cost_entries, replacing four independently-drifting direct-SQL write sites in dispatch.py, jobs.py (×2), and support_engineer.py. Upserts by job_id when present. Enforced by a call-site audit test that fails the suite if any other file writes to cost_entries directly.
• extract_tokens() now tags a .basis (regex_pair | total_split | none) distinguishing a real per-field token extraction from an 80/20 heuristic guess.
• .synlynk/model_rates.json (scaffolded by synlynk init) is now the source of truth for per-model rates and billing_mode, replacing a hardcoded table; falls back to hardcoded rates on missing/invalid file.
• New 3-tier t-shirt-size token fallback (_estimate_tshirt_tokens()) for surfaces with no real token count: story's estimated_tokens column → historical average from cost_entries (same discipline+phase, ≥3 samples) → fixed conservative default.
• synlynk cost log — manually log a cost row for native/unwrapped PM or brainstorming sessions with no CLI token data, tagged estimated_manual.
• Cost coverage closed across every surface identified in the design spec's audit: dispatch_agent()'s exec wrapper (no longer gated on in_tokens > 0), jobs.py's reconcile and daemon-reconcile paths, cmd_launch(), and support_engineer.py's investigation runs. synlynk release, synlynk probe, and synlynk doctor audited and confirmed correctly out of ledger scope.
• check_budgets() gained a dedicated sub-line surfacing failed-job placeholder estimates separately from the headline spend total. costs.md and budget parsers now tolerate [est] /[legacy] /~ prefixes.
• New "Cost Capture Protocol" section in CLAUDE.md.
• [failed job] marker fix: dispatch.py's zero-token-failure cost label now prepends (not appends) the marker so it survives update_costs()'s 20-character command truncation, restoring check_budgets()'s failed-job sub-line to live code.

Phase 2 — structured-output token adapters, one per dispatch CLI (PRs #244, #252, #256, #257): replaces the 80/20 heuristic split with a real per-field structured-JSON extraction path for each vendor CLI — Codex, Claude, Agy (Gemini), and Grok all shipped, closing epic #210's adapter scope.

Display layer (PRs #264, #266): Vizor's Effort & Cost tab now visually flags estimated-vs-measured cost rows; synlynk status gained a RATES line (and JSON rates_updated_at key) showing when the model rate table was last updated, warning ⚠ never updated when no rate file exists.

Design spec: docs/superpowers/specs/2026-07-13-measurement-ledger-hardening-design.md. Roadmap arc + epic #210 marked shipped (PR #267). Deferred follow-ups filed as issues, out of scope for this release: #260 (Savings Ledger), #261 (dispatch-path unification), #262 (surface consolidation), #263 (stage constants staleness).

Local Agent — 5th dispatch agent, zero-cost on-device inference (PRs #200/#204/#205/#207/#208/#209)

• local joins claude/codex/agy/grok as a dispatchable agent, running as an aider CLI subprocess against an on-device oMLX OpenAI-compatible endpoint — zero per-token cost.
• Capability envelope seeding + concurrency guard so local starts with a conservative starter whitelist that self-widens with verified results.
• Real-hardware opt-in pytest tier (tests/test_local_agent_hardware.py) exercises actual Aider+oMLX end-to-end, correctly skipped when no local hardware is running.
• Capability-matrix taxonomy worked example + blog post documenting the integration.

Capability Matrix Hardening — 3-stage routing engine + fleet batch scheduler (epic #137, PRs #139/#140/#141/#147/#148/#150/#151/#152/#154/#156)

• _best_agent_for_story() now scores candidates across three real stages: weighted capability score (VERIFIER tier, dead-signal handling, pr_review_cycles/verified_by_ci signals), a hard quota-headroom gate (agent_quotas table across 5h/hourly/daily/weekly/monthly windows, degraded-mode fallback when quota signal is missing), and a cost tie-break (cheaper model wins when top scores are within 0.15).
• New synlynk/scheduler.py: stories.priority/readiness columns, synlynk story ready/draft gate, fleet-level in-batch headroom accounting (story N in a batch sees story 1..N-1's projected spend), retry/reassignment capped at 2 attempts, synlynk schedule [--execute] [--max-stories N] CLI.
• GOVERNS seven-stage vocabulary rollout replaces the old CYCLES/CYCLE_COLOURS naming across the capability taxonomy and HUD.
• Capability tag enum enforcement + taxonomy reference doc; cycle_capability row dedup migration fix.
• Deferred v2 (reset-timing-aware bin-packing, persistent quota-blocking history, GOVERNS-aware readiness gate) tracked as a goal, gated on 30 days of v1 production data (deadline 2026-08-10) rather than scoped into stories now.

Job Lifecycle Ground-Truth Verification (#126, #127, #128, #129, PRs #130/#131/#132/#133/#135)

• dispatch_agent() creates a dedicated git worktree per dispatched job (worktrees/<job_id>, branch dispatch/<agent>/<job_id>) instead of sharing the invoking shell's cwd — concurrent dispatches no longer collide.
• _reconcile_jobs() cross-checks git state when a job's exit sentinel is missing, instead of treating an ambiguous exit as automatic failure; new "failed_unverified" status flags "inspect before discarding."
• files_touched is real (via git diff --name-only <merge-base> HEAD + git status --short --porcelain), no longer hardcoded to [].
• cmd_migrate() prints the resolved DB_PATH and fails loud with MigrationImportError on a 0-row import from a non-empty source, instead of a silent green banner.

Vizor

• Architect Map v2 (PR #167): replaces the static tube-map SVG with a live force-directed graph of workspace repos and typed cross-repo edges, a side drawer (path/stack/GitHub/dispatch/Gantt-jump/active-dream-count), and an IDE-style file-tree sub-view.
• Business Goals Panel (PR #153): surfaces goals/stories rollups in the HUD.
• __init__.py re-modularized a second pass — 11 focused modules extracted (synlynk/db.py, jobs.py, dispatch.py, context.py, sentinel.py, scheduler.py, etc.), no more single-file monolith (PR #180).

Fixed

Dispatch git-finalization reliability chain (#182/#184/#185/#189/#190/#191/#196/#198, PRs #186/#187/#192/#193/#194/#195/#199/#201)

• Dispatched agents (Codex/Agy) frequently completed real, tested work but didn't reliably finish their own git steps (commit/push/PR). synlynk now performs git finalization itself once, gated on a job's running→terminal transition (idempotent by construction) — stages everything except a hard-exclusion list, commits, pushes, opens a PR via gh pr create if none exists.
• Borrowed-worktree completions attributed via origin/<branch> state instead of being misread as zero-work.
• HARNESS_INTERNAL_TIMEOUT jobs now auto-retry (cap 2) instead of landing as a dead failure.
• A job that raced its own disk writes and landed failed with 0 touched is re-inspected once and reclassified failed_unverified rather than staying permanently stale.
• Cost accounting no longer bypasses the per-model rate table in 3 places (including a hardcoded gemini-2.5-pro $0 bug); local gets an explicit $0.0 override regardless of model version.
• Capability router now filters on canonical discipline, not the legacy engg_domain column.
• Daemon queue launch unified through dispatch_agent() so worktree/preflight/permission-flags/concurrency-guard logic isn't duplicated across two divergent paths.
• Harness no longer misreports completed jobs as FAILED with fabricated token counts.

Misc dispatch/CI hardening (PRs #163/#164/#165/#171/#173/#238/#240)

• Codex sandbox gets --add-dir <git-common-dir> so it can write git refs inside a worktree.
• dispatch --help's agent list now derives from AGENT_CAPABILITY_BASELINES instead of a stale hardcoded list.
• Stall-killer generalized to detect harness-internal timeouts and checks remote branch activity before hard-failing a silent job.
• 3 baseline CI flakes isolated from runner environment state; Python 3.8 compat fix for a tuple[str, str] annotation.
• dispatch --context-mode full now warns when used on a task that's already self-contained.

---

Release pitch: synlynk v0.11.0 is the Agent Ecosystem Operational Layer — every dispatch now carries permissions and recovery paths, your terminal shows a live fleet of agents in real time, and the full workflow discipline is baked into every agent directive file at init.

Added

Agent Autonomy Bridge (BS-12, PR #119)

• synlynk dispatch --grant <perm> --revoke <perm> — per-task permission overrides; role defaults in .synlynk/config.json map 12 roles (pm, review, implement, test, css, infra, etc.) to capability tiers. Resolved set translates to --allowedTools (Claude), --approval-policy (Codex), or a ## Permissions context header (Agy).
• synlynk configure agent <name> [--flag k=v] [--env K=V] [--network-dep host:port] — write per-project harness overrides to .agents/<agent>.json; dispatch_agent() merges at call time: baseline → per-project overrides → per-task grant/revoke.
• synlynk jobs --stalled — list jobs with HANDOFF_PENDING sentinel (set when a job accumulates STALL_NO_OUTPUT, FLATLINE, or QUOTA_EXHAUSTED). Shows job ID, agent, failure sentinel, elapsed time, and recommended next agent.
• synlynk jobs handoff <job_id> [--to <agent>] — transfer a stalled job to a new agent; appends ## Handoff Note to the job context file, increments handoff_count, updates previous_agents (JSON array), launches new dispatch with full context, clears HANDOFF_PENDING.
• synlynk doctor TC-5 — scans each directive file for all 6 required SOP section headers; warns (non-blocking) if any are absent. Interactive fix wizard runs after each TC failure with structured fix paths (TC-1–5) and an "I'm stuck" escape that assembles failure context and dispatches Claude for conversational diagnosis.
• synlynk sync --repair-sops — re-injects missing SOP sections into directive files without touching existing harness fence content. Idempotent.
• 6 SOP blocks written into CLAUDE.md, GEMINI.md, AGENTS.md, and GROK.md at synlynk init and synlynk sync time: PR Review Discipline · Brainstorm-First Policy · Design → Plan → Build Sequence · Capability-Based Task Allocation · Cost Visibility · Repo Hygiene.
• daemon_jobs schema additions: handoff_count INTEGER DEFAULT 0, previous_agents TEXT (JSON array). Migration applied automatically on first access.

Live Job Observatory (BS-13, PR #117)

• synlynk watch --live — fullscreen live job board: all running/recent jobs with agent, status, elapsed time, cost, and output tail; auto-refreshes every 3s; Ctrl-C to exit.
• Vizor Observatory tab — fifth tab in synlynk viz browser dashboard: real-time job fleet table (JS polling), per-agent status badges, cost rollup, job output drawer.

Vizor Efficiency Enrichment (BS-22, PRs #113, #118)

• R/W/T budget bars — per-agent card shows Read, Write, Test utilisation as percentage of TIER1_CAPACITY.
• Cycle × Agent capability matrix — 6 × 4 table (dream/plan/work/ship/maintain/engage × Claude/Agy/Codex/Grok) with full/partial/none badges (cap-full, cap-partial, cap-none CSS classes).
• Per-agent radar hexagon SVGs — 80×80px SVG radar on each agent card; 6-axis polygon filled at 30% opacity with agent theme colour; axis score maps full→1.0, partial→0.5, none→0.0.

Ecosystem Status + Capacity (BS-16, PR #110)

• synlynk status — terminal platform health: harness compliance table, agent availability, budget pulse (R/W/T capacity bars per agent), 6-cycle capability matrix.
• synlynk status --json — machine-readable ecosystem snapshot consumed by Vizor as its live data contract.
• Three dispatch modes: eco (respects R/W/T budget gates), daily-grind (default), perf (no budget gates).
• Three new _preflight_dispatch() gates: CAPACITY_EXCEEDED_INPUT, CAPACITY_EXCEEDED_OUTPUT, TOOL_PRESSURE.
• New state.db tables: harness_status, cycle_capability; probe command extended to seed Tier 1 capacity baselines.
• synlynk/status.py new module; HarnessSnapshot dataclass in hud.py.

Modularisation (chore, PRs #103–#109)

• Extracted synlynk/__init__.py (11,268L) into five focused modules: synlynk/probe.py, synlynk/sentinel.py, synlynk/upgrade.py, synlynk/dispatch.py, synlynk/_constants.py. __init__.py reduced to ~1,500L of orchestration and CLI surface.

Fixed

TC-2 dispatch fix arc (PRs #114–#116)

• Fixed false-positive TC-2 failures caused by Agy dispatch_flags baseline including --non-interactive (valid for Agy but flagged as invalid by harness validator). Corrected per-agent flag baseline maps.
• Fixed _run_tc2 seeding invalid flags into the passed list on first scan, blocking all subsequent Agy dispatches.
• synlynk logs now shows exit summary + TC-2 preflight gate result per dispatch.

BS-12 review fixes

• --repair-sops now merges missing SOP blocks into existing harness fence body instead of replacing the entire fence.
• TC-5 result wired into all_passed in cmd_doctor; doctor now exits non-zero when SOPs are missing.
• Handoff note included in handoff_task sent to new agent (was reading context before append).
• TC-5 fix menu scoped to agents with missing sections only (was firing for all agents in loop).
• --flag KEY (boolean flag without =) in configure agent now uses partition instead of crashing with ValueError.

BS-22 review fixes

• Removed is_placeholder short-circuit in SVG circles matrix that contradicted the new text capability matrix.
• CSS class cap-{support} now normalised to lowercase; prevents unstyled cells from mixed-case DB values.
• Removed unreachable dead-code fallback block in get_capability_level().

Changed

- `synlynk doctor` interactive fix wizard replaces silent print-and-exit after TC failures. - `_check_job_stall()` now also writes `HANDOFF_PENDING` sentinel on STALL_NO_OUTPUT / FLATLINE / QUOTA_EXHAUSTED. - `dispatch_agent()` merge layer: baseline → `.agents/.json` harness overrides → per-call `--grant`/`--revoke`.

---

v0.10.02026-07-03

Release pitch: synlynk v0.10.0 is the Developer Preview — install it via pipx, set up your workspace in 60 seconds with the terminal wizard, and get a live browser dashboard that shows exactly what your agent team is doing.

Added

FTUE + Onboarding

• synlynk init --wizard — FTUE typeform-style TUI wizard (6 screens: home harness detection, workspace topology, skills scan, agent fleet, role assignment, launch cheat sheet). Mandatory Phase 0 silent scan; writes workspace config, state.db, and role blocks into each agent's directive file. Ctrl-C before completion leaves no state.
• synlynk scan — re-runnable repo analysis: detects topology (single/mono/multi), fingerprints stack per repo/package via 14 file-presence heuristics, parses CLAUDE.md/GEMINI.md/AGENTS.md, maps to workspace in state.db, regenerates structured context.md. Flags: --refresh, --add <path>, --remove <path>, --dry-run.
• synlynk launch — FTUE task picker TUI with 6-cycle SDLC view (Dream · Plan · Work · Ship · Maintain · Engage); 12 scan-triggered launch templates (3 core + 9 stack-aware); dispatch preview screen; synlynk open replaces old synlynk launch <agent> for direct agent open. (BS-19, PR #94)
• synlynk roles — print current agent role table from .synlynk/config.json; synlynk init and synlynk doctor now generate per-agent directive role blocks (## Your Role in CLAUDE.md, GEMINI.md, AGENTS.md). (BS-12a, PR #95)

State + Migration

• synlynk migrate — one-shot atomic import of project-docs/ markdown into state.db (8 steps: import → copy to .synlynk/project-docs/ → git rm → .gitignore → sentinel → commit). After migration, every DB write immediately mirrors to .synlynk/project-docs/ as a local backup (write-through). Flags: --dry-run, --recover (re-import from backup after DB loss), --setup-dr (configure cloud-synced DR folder).
• synlynk memory add / synlynk devlog append — write memory entries and devlog sessions to state.db with immediate flat-file write-through.
• 5 new state.db tables: memory_entries, roadmap_arcs, roadmap_phases, cost_entries, devlog_entries; gh_issue column on stories.
• DR sync — configurable dr_sync_path in .synlynk/config.json; every write-through copy is also synced to a cloud-synced local folder (iCloud/GDrive/OneDrive). No OAuth, no new deps.

Deep Scan

• synlynk scan (deep mode) — 6-stage pipeline: repo fingerprint, dependency graph, test coverage ratio, doc coverage, CI health, churn density. Stage Cards TUI with progress indicators; scan fences written to state.db; synlynk launch templates upgrade automatically from scan signal data. 6 new scan fields: test_ratio, readme_word_count, has_ci, has_docs, has_type_hints, has_orm. (BS-20, PR #96)

Daily-Driver Commands

• synlynk jobs --summary <id> — after every job closes, print structured summary: files touched, exit status, cost, tokens, duration; append to .synlynk/logs/<job_id>.summary; readable via synlynk jobs --summary <id>. (PR #97)
• synlynk release — Ship cycle stub: bump VERSION, generate CHANGELOG entry from merged stories since last tag, write blog post stub in docs/blog/; --dry-run previews without writing. (PR #98)
• synlynk status --platform — infrastructure health view: harness compliance (last probe, any DRIFT sentinels), agent availability table (installed/version/TC status), budget pulse (daily/weekly burn rate). (PR #99)

Browser Dashboard

• synlynk viz — 5-view local browser dashboard generated from state.db, served at http://localhost:8721. Views: Gantt (accordion drill-down, stage bars, pencil notes), User Journeys (split-pane, docs/journeys/*.md), Architect Map (tube map SVG from vizor-tube.json), Effort & Cost (SVG bar charts), Efficiency (agent report cards + sentinel timeline). Sticky note system: POST /note → viz-notes.json → injected into generate_context() (visual annotation → AI context loop). Live JS polling + browser notifications. Zero new deps. (BS-21, PR #101)

Packaging

• pipx packaging — pyproject.toml with [project.scripts] entry point; VERSION is the single source of truth in synlynk/__init__.py (pyproject.toml reads it via dynamic attr). Install via pipx install git+https://github.com/nikhilsoman/synlynk.
• _detect_install_type() — detects pipx vs pip vs script install; synlynk upgrade routes to pipx upgrade synlynk when installed via pipx.

Changed

- **`generate_context()`** routes to `_generate_context_from_db()` when `.synlynk/.synlynk_migrated` sentinel is present; reads from state.db (top story, recent devlog entries, recent memory sections). - **`install.sh`** derives VERSION dynamically from `synlynk/__init__.py` instead of hardcoding. - **Python requirement** raised from 3.8+ to 3.9+. - **Refactor:** `main()` extracted to `synlynk/cli.py`; data-layer functions extracted to `synlynk/db.py`. Single-file `bin/synlynk.py` is now a thin dispatcher. - README fully overhauled: pipx install, badge strip, wizard-first 60-second quickstart, state.db architecture, all new commands documented.

Tests

• 747 tests (up from 588 at v0.9.8); 28 new migrate tests, 28 launch/FTUE tests, 21 Vizor tests, 7 packaging tests, 40 deep-scan tests, full E2E round-trips.

---

v0.9.82026-06-27

Added

- **`synlynk exit`** — reverse all synlynk onboarding: strips managed sections from tracked instruction files (CLAUDE.md, GEMINI.md, etc.), removes `.agents/` profiles and `.synlynk/` directory, writes `SYNLYNK_HANDOFF.md` with re-init instructions. Dry-run by default; `--confirm` to execute. `--remove-docs` optionally removes `project-docs/`. - **`synlynk repair`** — exit + immediate re-init from captured config (agents, mode, org, repo, docs-dir). Dry-run by default; `--confirm` to execute. - **`synlynk sync`** — propagate updated synlynk artifacts (instruction file sections, missing `.agents/` profiles) to an existing repo without full re-init. Dry-run by default; `--confirm` to execute. - **`_strip_synlynk_section(path, marker_style)`** internal helper — removes synlynk-managed block from any instruction file; handles html/hash/none marker styles; leaves surrounding user content intact.

Changed

- VERSION bumped `0.9.7 → 0.9.8`

---

v0.9.72026-06-26

Added

- **Grok as a first-class fourth agent peer** alongside claude/agy/codex across all synlynk subsystems - `AGENT_CAPABILITY_BASELINES["grok"]` — cli, non_interactive_flags (`-p`), prompt_via_arg, dispatch_flags (`--always-approve`), roles (builder/architect), strengths - `AGENT_DISCOVERY_DEFAULTS["grok"]` — discovery path `~/.grok` - `_probe_model_version` — `grok -v` probe + `grok-[\w.-]+` version pattern - `GROK.md` template — identity (`Co-Authored-By: Grok `), branch prefixes (`feat/grok/`, `fix/grok/`), standard session/worktree/live-issues sections, `synlynk:start/end` markers - `_INSTRUCTION_TARGETS` + `_MARKER_STYLE_FOR_TOOL` entries for GROK.md - Init wizard: GROK.md in `trio_content` + `_agent_guards`; `agent_slots` default expanded to four agents; `--agents` default updated to `claude,agy,codex,grok` - `_inject_grok_rules()` — prepends `--rules GROK.md` for all grok exec calls; adds `--rules .synlynk/context.md` in headless (`-p`) mode; silently skips missing files - `dispatch_agent()` — `--always-approve` → `--permission-mode bypassPermissions` fallback via agent profile `always_approve_unsupported`; `--output-format json` for grok headless dispatch - `extract_tokens()` — nested `usage.input_tokens/output_tokens` pattern for Grok JSON output - `extract_model_version()` — tier-2 path via `.agents/grok.json` `"model"` field - `GROK.md` written to the synlynk repo itself (100 lines, markers bookending) - 15 new tests covering all registration, instruction file, init wizard, exec injection, dispatch, and token/model extraction paths

Fixed

- Stale time-sensitive fixture in `test_collect_capability_drop_returns_finding` — hardcoded 2026-06-21 timestamps replaced with `datetime.now(timezone.utc)`-relative values

---

v0.9.32026-06-23

Added

- `SynlynkDaemon` class — subclasses `WatchDaemon` to add an embedded HTTP server thread on `localhost:27471` and persistent job dispatch on every poll tick; double-fork daemonization inherited from `WatchDaemon`; separate pidfile `.synlynk/daemon.pid` and log `.synlynk/daemon.log` - `daemon_jobs` table in `state.db` — persistent job queue with `priority` (1–10), `depends_on` (JSON array of job IDs), and full status lifecycle `queued → running → done | failed` - `_reconcile_daemon_jobs()` — reaps finished child processes using `os.waitpid(WNOHANG)` (zombie-safe), reads `.exit` files for exit codes, updates `status`/`exit_code`/`completed_at` in state.db - `_dispatch_ready_jobs(max_parallel)` — launches queued jobs respecting concurrency cap and dependency chains; propagates `failed` status to downstream dependents immediately; commits per-job for crash-safe restart semantics - HTTP API on `localhost:27471` — 10 endpoints: `GET /context`, `GET /status`, `GET /jobs`, `GET /jobs/`, `POST /dispatch`, `GET /stories`, `GET /stories/`, `GET /capability`, `GET /sentinel`, `POST /checkpoint`; `_ReuseAddrHTTPServer` subclass prevents `Address already in use` on rapid restart; `threading.Lock` guards concurrent `generate_context()` calls - `synlynk daemon start|stop|status|restart` CLI command - `synlynk daemon --install-service` — writes and activates platform service unit: macOS: `~/Library/LaunchAgents/com.synlynk.daemon.plist` via `launchctl load -w`; Linux: `~/.config/systemd/user/synlynk-daemon.service` via `systemctl --user enable --now`; Fallback: `@reboot` crontab entry - `synlynk daemon --uninstall-service` — reverses each platform path; handles `FileNotFoundError` gracefully

Fixed

- Daemon zombie process detection: replaced `os.kill(pid, 0)` with `os.waitpid(WNOHANG)` so exited child processes are properly reaped rather than staying `running` indefinitely - Dependency deadlock: queued jobs whose dependency fails are immediately marked `failed` rather than staying queued forever - Transaction isolation in dispatch: each launched job is committed individually so a crash mid-loop cannot produce duplicate spawns on restart

---

v0.9.22026-06-22

Added

- `synlynk join` — new member onboarding: seeds a devlog stub for the joining user, regenerates AI context files (CLAUDE.md, GEMINI.md, AGENTS.md) with the joining member's identity, and prints a team digest showing all active members and their recent focus areas - `synlynk team status` — team digest view: lists all members with devlog presence, current story assignments, token budget consumption, and last-active timestamp; reads `project-docs/devlogs/.md` across all contributors - `synlynk decide --panel ` — multi-agent consensus panel: dispatches the panel agents non-interactively with the same decision prompt, collects structured `DECISION:` blocks from each response, computes a consensus position, and optionally writes a signed `Decision` record to `project-docs/decisions/YYYY-MM-DD-.md` with `--record` flag - `--tokens ` flag on `synlynk story create` — set an estimated token budget for a story; stored in new `estimated_tokens` column on the `stories` table - `_seed_devlog(username, root)` helper — writes a devlog stub for the joining user with an initial entry so the devlog file exists and is attributable in team digests from day one - `_generate_ai_context_files(username, root, config)` helper — regenerates CLAUDE.md, GEMINI.md, and AGENTS.md with the joining member's name in the `git config user.name` slot - `_build_team_digest(root)` helper — reads all devlogs from `project-docs/devlogs/`, extracts last-active date and focus summary per member; used by both `join` and `team status` - `_check_upstream_divergence(root)` helper — checks whether the local branch is behind the remote before any write to `project-docs/`; prints a warning with advice to `git pull` if divergence is detected; continues without blocking so offline workflows still function (pull-before-write arbitration)

Fixed

- None in this release

---

v0.9.12026-06-22

Added

- `--docs-dir ` flag on `synlynk init` — override the default `project-docs/` location for repos that store docs elsewhere (e.g. `--docs-dir docs/project`) - `_docs_dir(root, config)` helper — reads `docs_dir` from `.synlynk/config.json`, falls back to `project-docs/`; used by `exec`, `checkpoint`, `status`, and `init` so every command respects the configured docs location

Fixed

- Installed binary (`~/.synlynk/bin/synlynk`) crashed with `ModuleNotFoundError: No module named 'synlynk'` when invoked outside the source repo after the v0.9.0 package split. Fixed by embedding the package `synlynk/` directory in `~/.synlynk/lib/synlynk/` at install time and prepending `$HOME/.synlynk/lib` to `sys.path` in the installed shim. - `synlynk init` in a repo with existing `project-docs/` overwrote the existing docs with blank templates. Fixed by detecting existing files via `_find_existing_doc()` and migrating their content to the new location (or skipping write if no relocation is needed).

---

v0.9.02026-06-21

Added

- Scoped dispatch context: `exec` now injects a per-task section (`## Current Task`) with only the relevant plan block instead of the full devlog, reducing context window usage - `## Relevant Files` injected per dispatch from the source map — derived from the task description matched against `source-map.md` symbols - `## How to Verify` contract injected per dispatch — specifies acceptance criteria the agent should validate before declaring the task done - Per-agent prompt framing: Claude, AGY, and Codex each receive a tailored preamble that matches their CLI interaction model (conversational vs. task-oriented vs. non-interactive) - Ed25519 capability rating signing: every `capability_ratings` row is signed with the project key so ratings cannot be forged across project boundaries - Anti-gaming quality cap: `test_count < 3` stories are capped at quality score 5.0 regardless of other signals, preventing artificially inflated ratings for untested work - `synlynk/` package split: all ~5000 lines of application logic moved from `bin/synlynk.py` into `synlynk/__init__.py`; `bin/synlynk.py` becomes a 5-line import shim

Fixed

- `capability_ratings` entries were not being attributed to the correct project when multiple synlynk-managed repos shared the same `~/.synlynk/` directory — Ed25519 project key now scopes all ratings correctly

---

v0.8.02026-06-21

Added

- `synlynk agent run` — foreground support engineer investigation: collects signals, formats a structured report, and offers to file a GitHub issue or draft a fix PR - Five signal collectors for the support engineer archetype: failing tests, flaky tests, coverage gaps, stale dependencies, open GitHub issues exceeding age threshold - 7-day and 30-day deduplication: signals already filed within the window are suppressed so the agent doesn't file duplicate issues on repeated runs - `synlynk agent --install-cron` — registers a launchd plist (macOS) or systemd timer (Linux) that runs the support engineer agent on a configurable schedule - `.agents/` config directory: `support-engineer.json` defines signal weights, age thresholds, and notification channels per project; read by `synlynk agent run` at startup - GitHub issue filing via `gh issue create` with structured body including signal summary, affected files, and suggested fix skeleton - Draft fix PR creation via `gh pr create` for issues with high-confidence fix candidates

---

v0.7.02026-06-20

Added

- `synlynk scan` / `synlynk scan --deep` — language-agnostic source scanner: reads file tree, extracts top-level symbols from Python/JS/TS/Go/Rust/Ruby/Java/C/C++ source, writes `source-map.md` and populates `source_symbols` table in `state.db` - `## Source Architecture` section injected into every `exec` context from the cached scan result, giving agents a structural overview without reading individual files - Passive git-HEAD cache: scan results are keyed to the current git HEAD SHA; a re-scan is only triggered when HEAD changes, not on every `exec` call - `synlynk scan --status` — shows last scan timestamp, HEAD SHA, and symbol count without re-scanning - Dual storage: symbols written to both SQLite `source_symbols` table (queryable) and `source-map.md` (human-readable, injected into agent context) - Language detection by file extension with fallback to content heuristics

Fixed

- `synlynk scan` no longer traverses `.git/`, `node_modules/`, or `.synlynk/` directories

---

v0.6.12026-06-17

Added

- Instruction reach to seven additional IDE/editor targets: Cursor (`.cursor/rules/`), GitHub Copilot (`.github/copilot-instructions.md`), Windsurf (`.windsurfrules`), Cline (`.clinerules`), Aider (`.aider.conf.yml`), Continue (`.continue/config.json`), and Sourcegraph (`.sourcegraph/memory.md`) - SHA manifest (`instructions.json`) tracking the synlynk-managed section hash for each generated file; used for drift detection - Runtime drift detection: `exec` warns if any tracked instruction file's section has been externally modified since last generation - `synlynk instructions status / diff / update / ack` — manage instruction file state - Task status model: 5 states (`active`, `done`, `deferred`, `superseded`, `absorbed`); deferred tasks are included in context with reduced weight; `checkpoint` archives resolved states to a separate section - AGY CLI replaces Gemini CLI throughout: all references to `gemini` updated to `agy` - VERSION synced to GitHub releases (was incorrectly stuck at 0.4.x) - `DB_PATH` centralised to `~/.synlynk/projects//state.db` so the database is shared across worktrees of the same project

---

v0.6.02026-06-14

Added

- Model version tier-2 probe: `discover_agents()` now probes for Opus/Sonnet/Pro variants in addition to the base model, annotates capability entries with `model_tier` - `synlynk pr check` — validates that the current branch's diff satisfies the story's acceptance criteria before opening a PR; exits non-zero if criteria unmet - `synlynk score attest` — manually attest a story's quality score with a signed reason; appended to `capability_ratings` with `attestation=true` flag - Verifier pipeline output capture: `run --trio` now captures and surfaces the Verifier agent's structured review comment - Tokq `org_domain_tags` capability dimension: stories can be tagged with domain taxonomy labels (`backend/api`, `frontend/ui`, etc.) for cross-project capability aggregation - Constraint propagation: blocking story constraints propagate to child tasks; dispatching a child task for a blocked story emits a warning and requires `--force`

---

v0.5.02026-06-14

Added

- SQLite WAL state database (`~/.synlynk/projects//state.db`) replacing the flat JSON job store; WAL mode enables concurrent reads from multiple agent processes - Model-aware routing: `dispatch` selects agent by matching story domain tags against `capability_ratings`; no domain match falls back to round-robin - 3D domain taxonomy for capability rating: `(agent, domain, model_tier)` composite key replaces flat per-agent scores - Quality signal hierarchy: test coverage, PR review comments, story completion rate, and task duration all feed the capability score; weights configurable in `.synlynk/config.json` - `synlynk story create