opencode-harness-kit 0.0.0-stage → 0.2.0

package/profiles/deepseek-v4-flash-free.json

@@ -0,0 +1,41 @@
+{
+  "$comment": "Fallback tier 3. The smallest context in the queue at 200k, which is why prune.keepRecentToolCalls drops: the same settings tuned for a 1M model overflow this one.",
+  "id": "deepseek-v4-flash-free",
+  "title": "DeepSeek V4 Flash Free (OpenCode Zen)",
+  "provider": "opencode",
+  "model": "deepseek-v4-flash-free",
+  "family": "deepseek-flash",
+  "verified": "2026-10-07",
+  "retention": "zero-retention",
+  "notes": "Text only, no attachment, 200k context. Included to show the compiler scaling pruning to the model: reserveShare and retention fall together, so a 1M-tuned profile cannot silently overflow a 200k model.",
+  "limits": {
+    "context": 200000,
+    "output": 128000
+  },
+  "supports": {
+    "reasoning": true,
+    "temperature": true,
+    "toolCall": true,
+    "structuredOutput": true,
+    "attachment": false
+  },
+  "effort": {
+    "levels": ["low", "high", "max"],
+    "default": "max",
+    "reserveShare": 0.12
+  },
+  "roles": {
+    "plan": "max",
+    "build": "high",
+    "reviewer": "max",
+    "locate": "low",
+    "verify": "low"
+  },
+  "prune": {
+    "enabled": true,
+    "keepRecentToolCalls": 4,
+    "keepRecentTurns": 2,
+    "keepUserMessages": true
+  },
+  "fallback": []
+}

package/profiles/index.json

@@ -0,0 +1,78 @@
+{
+  "$comment": "Profiles verified against live provider metadata. `retention` must come from retention.yml; a model missing there cannot enter this list.",
+  "default": "space-bunny-free",
+  "profiles": [
+    {
+      "id": "muse-split-free",
+      "file": "profiles/muse-split-free.json",
+      "title": "Muse Spark 1.3 (heavy) + Space Bunny Free (bulk output)",
+      "retention": "may-train",
+      "free": true,
+      "topEffort": "max",
+      "context": 1048576,
+      "default": false,
+      "split": {
+        "plan": "muse-spark-1.3-contributor-free",
+        "build": "muse-spark-1.3-contributor-free",
+        "reviewer": "muse-spark-1.3-contributor-free",
+        "locate": "space-bunny-free",
+        "verify": "space-bunny-free"
+      },
+      "note": "Retention here is the effective one: three roles run on Muse Spark, so the file is may-train even though the base model is zero-retention."
+    },
+    {
+      "id": "space-bunny-free",
+      "file": "profiles/space-bunny-free.json",
+      "title": "Space Bunny Free",
+      "retention": "zero-retention",
+      "free": true,
+      "topEffort": "max",
+      "context": 1048576,
+      "default": true,
+      "note": "Only free Zen model with a max effort and zero-retention terms."
+    },
+    {
+      "id": "x-preview-f-free",
+      "file": "profiles/x-preview-f-free.json",
+      "title": "x-preview-f Free",
+      "retention": "zero-retention",
+      "free": true,
+      "topEffort": "max",
+      "context": 1000000,
+      "default": false,
+      "note": "Three effort levels only; no medium or xhigh."
+    },
+    {
+      "id": "longcat-2.5-preview-free",
+      "file": "profiles/longcat-2.5-preview-free.json",
+      "title": "LongCat 2.5 Preview Free",
+      "retention": "zero-retention",
+      "free": true,
+      "topEffort": null,
+      "context": 1000000,
+      "default": false,
+      "note": "No tunable reasoning effort."
+    },
+    {
+      "id": "deepseek-v4-flash-free",
+      "file": "profiles/deepseek-v4-flash-free.json",
+      "title": "DeepSeek V4 Flash Free",
+      "retention": "zero-retention",
+      "free": true,
+      "topEffort": "max",
+      "context": 200000,
+      "default": false,
+      "note": "200k context, text only."
+    }
+  ],
+  "excluded": [
+    {
+      "id": "muse-spark-1.3-contributor-free",
+      "file": "examples/muse-spark-1.3-contributor-free.json",
+      "title": "Muse Spark 1.3 Contributor Free",
+      "retention": "may-train",
+      "reason": "Provider may train on your prompts and completions (Meta contributor tier). Top effort is xhigh, not max.",
+      "installable": true
+    }
+  ]
+}

package/profiles/longcat-2.5-preview-free.json

@@ -0,0 +1,41 @@
+{
+  "$comment": "Fallback tier 2. No tunable reasoning effort at all, so the loader omits `variant` entirely for this profile rather than guessing a level.",
+  "id": "longcat-2.5-preview-free",
+  "title": "LongCat 2.5 Preview Free (OpenCode Zen)",
+  "provider": "opencode",
+  "model": "longcat-2.5-preview-free",
+  "family": "longcat",
+  "verified": "2026-10-07",
+  "retention": "zero-retention",
+  "notes": "Zero-retention with a 1M context, but reasoning_options is null upstream: the model has one fixed effort. Any role requesting a level resolves to null and the generated config leaves variant unset.",
+  "limits": {
+    "context": 1000000,
+    "output": 131072
+  },
+  "supports": {
+    "reasoning": true,
+    "temperature": true,
+    "toolCall": true,
+    "structuredOutput": false,
+    "attachment": true
+  },
+  "effort": {
+    "levels": [],
+    "default": null,
+    "reserveShare": 0.12
+  },
+  "roles": {
+    "plan": null,
+    "build": null,
+    "reviewer": null,
+    "locate": null,
+    "verify": null
+  },
+  "prune": {
+    "enabled": true,
+    "keepRecentToolCalls": 6,
+    "keepRecentTurns": 3,
+    "keepUserMessages": true
+  },
+  "fallback": ["deepseek-v4-flash-free"]
+}

package/profiles/muse-split-free.json

@@ -0,0 +1,64 @@
+{
+  "$comment": "Split profile. The heavy reasoning roles run on Muse Spark 1.3 because it is the strongest free model on Zen by a wide margin (Terminal-Bench 2.1 88.8% vs Nemotron 3 Ultra's 56.4%, DeepSWE v1.1 75.4%). The bulk-output roles stay on Space Bunny because its 524,288-token output cap is 4x Muse Spark's 131,072, and scaffolding or migrating many files is where that ceiling is actually felt. `retention` below describes the BASE model only; the config header prints the effective retention, which is may-train because three roles run on Muse Spark.",
+  "id": "muse-split-free",
+  "title": "Muse Spark 1.3 (heavy) + Space Bunny Free (bulk output)",
+  "provider": "opencode",
+  "model": "space-bunny-free",
+  "family": null,
+  "verified": "2026-10-07",
+  "retention": "zero-retention",
+  "notes": "Opt-in split, not a default. Muse Spark is a Meta contributor-tier free endpoint: its provider may train on your prompts and completions. Three of the five roles are therefore may-train even though the base model is zero-retention. Install it only for public or throwaway repositories. Effort requests below are resolved against each role's own model, so plan and reviewer ask for `max` and land on `xhigh`, which is the top Muse Spark offers.",
+  "limits": {
+    "context": 1048576,
+    "input": 524288,
+    "output": 524288
+  },
+  "supports": {
+    "reasoning": true,
+    "temperature": false,
+    "toolCall": true,
+    "structuredOutput": false,
+    "attachment": true
+  },
+  "effort": {
+    "levels": ["low", "medium", "high", "xhigh", "max"],
+    "default": "max",
+    "reserveShare": 0.12
+  },
+  "roles": {
+    "plan": "max",
+    "build": "high",
+    "reviewer": "max",
+    "locate": "low",
+    "verify": "low"
+  },
+  "roleModels": {
+    "plan": {
+      "model": "muse-spark-1.3-contributor-free",
+      "levels": ["minimal", "low", "medium", "high", "xhigh"]
+    },
+    "build": {
+      "model": "muse-spark-1.3-contributor-free",
+      "levels": ["minimal", "low", "medium", "high", "xhigh"]
+    },
+    "reviewer": {
+      "model": "muse-spark-1.3-contributor-free",
+      "levels": ["minimal", "low", "medium", "high", "xhigh"]
+    },
+    "locate": {
+      "model": "space-bunny-free",
+      "levels": ["low", "medium", "high", "xhigh", "max"]
+    },
+    "verify": {
+      "model": "space-bunny-free",
+      "levels": ["low", "medium", "high", "xhigh", "max"]
+    }
+  },
+  "prune": {
+    "enabled": true,
+    "keepRecentToolCalls": 6,
+    "keepRecentTurns": 3,
+    "keepUserMessages": true
+  },
+  "fallback": ["x-preview-f-free", "longcat-2.5-preview-free", "deepseek-v4-flash-free"]
+}

package/profiles/schema.json

@@ -0,0 +1,161 @@
+{
+  "$schema": "http://json-schema.org/draft-07/schema#",
+  "$id": "https://github.com/anomalyco/opencode-harness-kit/profiles/schema.json",
+  "title": "OpenCode harness profile",
+  "description": "Everything that changes when you change model. Plugins read this file and contain no model names of their own.",
+  "type": "object",
+  "required": ["id", "provider", "model", "retention", "limits", "effort", "roles"],
+  "additionalProperties": false,
+  "properties": {
+    "id": {
+      "type": "string",
+      "description": "Slug used as the profile filename and the oc-harness use argument."
+    },
+    "title": { "type": "string" },
+    "provider": {
+      "type": "string",
+      "description": "opencode provider id, e.g. opencode, opencode-go, anthropic."
+    },
+    "model": {
+      "type": "string",
+      "description": "Model id without the provider prefix, e.g. space-bunny-free."
+    },
+    "family": {
+      "type": ["string", "null"],
+      "description": "Model family for grouping. Stealth models report null upstream, so this may be set by hand."
+    },
+    "verified": {
+      "type": "string",
+      "description": "ISO date a human checked this profile against live metadata."
+    },
+    "retention": {
+      "enum": ["zero-retention", "may-train", "unknown"],
+      "description": "Must match an entry in retention.yml. A profile whose model is unknown there cannot become the default."
+    },
+    "notes": { "type": "string" },
+    "limits": {
+      "type": "object",
+      "required": ["context", "output"],
+      "additionalProperties": false,
+      "description": "Token ceilings from the provider catalog. `input` is absent for several Zen models, in which case the loader derives it from `context`.",
+      "properties": {
+        "context": { "type": "integer", "minimum": 1 },
+        "input": { "type": "integer", "minimum": 1 },
+        "output": { "type": "integer", "minimum": 1 }
+      }
+    },
+    "supports": {
+      "type": "object",
+      "additionalProperties": false,
+      "description": "Capability flags. Missing or false means do not emit the related config.",
+      "properties": {
+        "reasoning": { "type": "boolean" },
+        "temperature": { "type": "boolean" },
+        "toolCall": { "type": "boolean" },
+        "structuredOutput": { "type": "boolean" },
+        "attachment": {
+          "type": "boolean",
+          "description": "Accepts image or video input. Gates screenshot-based verification."
+        }
+      }
+    },
+    "effort": {
+      "type": "object",
+      "description": "Available reasoning levels, lowest to highest. Empty means the model has no tunable effort.",
+      "properties": {
+        "levels": {
+          "type": "array",
+          "items": { "type": "string" },
+          "examples": [["low", "medium", "high", "xhigh", "max"]]
+        },
+        "default": { "type": "string" },
+        "reserveShare": {
+          "type": "number",
+          "minimum": 0,
+          "maximum": 0.5,
+          "default": 0.12,
+          "description": "Share of the input ceiling held back for reasoning and tool output growth."
+        }
+      }
+    },
+    "roles": {
+      "type": "object",
+      "description": "Requested effort per agent role. Resolved against effort.levels, never emitted raw.",
+      "additionalProperties": false,
+      "properties": {
+        "plan": { "$ref": "#/definitions/effortRequest" },
+        "build": { "$ref": "#/definitions/effortRequest" },
+        "reviewer": { "$ref": "#/definitions/effortRequest" },
+        "locate": { "$ref": "#/definitions/effortRequest" },
+        "verify": { "$ref": "#/definitions/effortRequest" }
+      }
+    },
+    "roleModels": {
+      "type": "object",
+      "description": "Model per agent role, for a profile that splits the heavy reasoning roles onto a stronger model while the bulk-output roles stay on a high-output one. `model` remains the base: roles not named here, the top-level `model`, and `small_model` all resolve to it.",
+      "additionalProperties": false,
+      "properties": {
+        "plan": { "$ref": "#/definitions/roleModel" },
+        "build": { "$ref": "#/definitions/roleModel" },
+        "reviewer": { "$ref": "#/definitions/roleModel" },
+        "locate": { "$ref": "#/definitions/roleModel" },
+        "verify": { "$ref": "#/definitions/roleModel" }
+      }
+    },
+    "prune": {
+      "type": "object",
+      "additionalProperties": false,
+      "description": "Deterministic context pruning. Only used when a model-driven compressor is unavailable or unproven.",
+      "properties": {
+        "enabled": { "type": "boolean" },
+        "keepRecentToolCalls": {
+          "type": "integer",
+          "minimum": 1,
+          "default": 6,
+          "description": "Verbatim tool outputs kept at the tail."
+        },
+        "keepRecentTurns": {
+          "type": "integer",
+          "minimum": 1,
+          "default": 3,
+          "description": "Trailing turns whose text and reasoning stay verbatim."
+        },
+        "keepUserMessages": {
+          "type": "boolean",
+          "default": true,
+          "description": "Never stub a user message. Their instructions are the point of the session."
+        }
+      }
+    },
+    "fallback": {
+      "type": "array",
+      "description": "Ordered profile ids to fall back to. Selection happens at session start, not mid-session: no opencode plugin hook can change a live session's model.",
+      "items": { "type": "string" }
+    }
+  },
+  "definitions": {
+    "roleModel": {
+      "type": "object",
+      "description": "A model for one role, plus the effort levels that model actually offers. Required when the model differs from the base, because a level the model does not have must never be written onto the agent.",
+      "additionalProperties": false,
+      "required": ["model"],
+      "properties": {
+        "model": {
+          "type": "string",
+          "description": "Model id without the provider prefix. Must resolve to a profile in this directory, so its limits and retention are known."
+        },
+        "levels": {
+          "type": "array",
+          "items": { "type": "string" },
+          "description": "Effort levels this role's model offers. Omitted means the role emits no variant at all, which is the right reading for a model with no tunable effort."
+        }
+      }
+    },
+    "effortRequest": {
+      "oneOf": [
+        { "type": "string", "description": "Desired level, resolved to the nearest available one." },
+        { "type": "null", "description": "This role should not set a variant at all." }
+      ]
+    }
+  }
+}

package/profiles/space-bunny-free.json

@@ -0,0 +1,42 @@
+{
+  "$comment": "Validated against models.dev on 2026-10-07. Regenerate with `oc-harness use --refresh`.",
+  "id": "space-bunny-free",
+  "title": "Space Bunny Free (OpenCode Zen)",
+  "provider": "opencode",
+  "model": "space-bunny-free",
+  "family": null,
+  "verified": "2026-10-07",
+  "retention": "zero-retention",
+  "notes": "Stealth preview reasoning model. Served over /v1/chat/completions (openai-compatible), which is the weaker cache dialect compared with models on /v1/responses. At the time of writing it is the only free model on Zen whose top reasoning effort is `max` and whose provider is zero-retention. Free for a limited time; expect the slot to rotate.",
+  "limits": {
+    "context": 1048576,
+    "input": 524288,
+    "output": 524288
+  },
+  "supports": {
+    "reasoning": true,
+    "temperature": false,
+    "toolCall": true,
+    "structuredOutput": false,
+    "attachment": true
+  },
+  "effort": {
+    "levels": ["low", "medium", "high", "xhigh", "max"],
+    "default": "max",
+    "reserveShare": 0.12
+  },
+  "roles": {
+    "plan": "max",
+    "build": "high",
+    "reviewer": "max",
+    "locate": "low",
+    "verify": "low"
+  },
+  "prune": {
+    "enabled": true,
+    "keepRecentToolCalls": 6,
+    "keepRecentTurns": 3,
+    "keepUserMessages": true
+  },
+  "fallback": ["x-preview-f-free", "longcat-2.5-preview-free", "deepseek-v4-flash-free"]
+}

package/profiles/x-preview-f-free.json

@@ -0,0 +1,41 @@
+{
+  "$comment": "Fallback tier 1 for space-bunny-free. Zero-retention, 1M context, and a `max` effort, but only three levels: there is no medium or xhigh, so roles asking for those resolve down to high.",
+  "id": "x-preview-f-free",
+  "title": "x-preview-f Free (OpenCode Zen)",
+  "provider": "opencode",
+  "model": "x-preview-f-free",
+  "family": null,
+  "verified": "2026-10-07",
+  "retention": "zero-retention",
+  "notes": "Only three effort levels (low, high, max) and a 131k completion ceiling, so long single answers are not its strength. Kept as first fallback: same 1M context class and the same zero-retention terms as the default.",
+  "limits": {
+    "context": 1000000,
+    "output": 131072
+  },
+  "supports": {
+    "reasoning": true,
+    "temperature": true,
+    "toolCall": true,
+    "structuredOutput": true,
+    "attachment": true
+  },
+  "effort": {
+    "levels": ["low", "high", "max"],
+    "default": "max",
+    "reserveShare": 0.12
+  },
+  "roles": {
+    "plan": "max",
+    "build": "high",
+    "reviewer": "max",
+    "locate": "low",
+    "verify": "low"
+  },
+  "prune": {
+    "enabled": true,
+    "keepRecentToolCalls": 6,
+    "keepRecentTurns": 3,
+    "keepUserMessages": true
+  },
+  "fallback": ["longcat-2.5-preview-free", "deepseek-v4-flash-free"]
+}

package/retention.yml

@@ -0,0 +1,110 @@
+# Data retention and privacy, curated by hand.
+#
+# models.dev describes capability and price. It does NOT describe what a
+# provider does with your prompts. That gap is why this file exists: the most
+# consequential choice in a coding agent is often invisible in the metadata.
+#
+# Values:
+#   zero-retention  provider keeps nothing, no training
+#   may-train       prompts and/or completions may be used to improve the model
+#   unknown         not verified here; `oc-harness use` refuses to make it a
+#                   default until a human fills this in
+#
+# Sources: https://opencode.ai/docs/zen (Privacy section). Re-verify on their
+# release notes; terms change without a model rename.
+
+defaults:
+  unknown: unknown
+
+models:
+  opencode/space-bunny-free:
+    retention: zero-retention
+    note: >-
+      "Its provider follows a zero-retention policy and does not use your data
+      for model training." Free for a limited time; stealth model with no public
+      identity.
+    verified: 2026-10-07
+    source: https://opencode.ai/docs/zen
+
+  opencode/space-bunny:
+    retention: zero-retention
+    note: Same provider as space-bunny-free, paid route on the Go plan.
+    verified: 2026-10-07
+    source: https://opencode.ai/docs/zen
+
+  opencode/longcat-2.5-preview-free:
+    retention: zero-retention
+    note: "free on OpenCode for a limited time. Its provider follows a zero-retention policy."
+    verified: 2026-10-07
+    source: https://opencode.ai/docs/zen
+
+  # The two entries below are not named individually in the Privacy section.
+  # They fall under the general policy - providers follow zero-retention and do
+  # not train on your data, with a named exception list - so this is inferred
+  # from that list rather than quoted. Re-check before relying on it for
+  # anything sensitive.
+
+  opencode/x-preview-f-free:
+    retention: zero-retention
+    note: >-
+      Not named in the Privacy exception list, so the general zero-retention
+      policy applies. Inferred, not quoted.
+    verified: 2026-10-07
+    source: https://opencode.ai/docs/zen
+
+  opencode/deepseek-v4-flash-free:
+    retention: zero-retention
+    note: >-
+      Not named in the Privacy exception list, so the general zero-retention
+      policy applies. Inferred, not quoted. This route is also documented as
+      returning 429 for extended periods, so see docs/KNOWN-LIMITS.md.
+    verified: 2026-10-07
+    source: https://opencode.ai/docs/zen
+
+  opencode/muse-spark-1.3-contributor-free:
+    retention: may-train
+    note: >-
+      "Heavily discounted token pricing in exchange for permission to use your
+      prompts and completions to train future Meta models." 1M context, but its
+      top reasoning effort is xhigh, not max. Kept as an example of a
+      capability/privacy tradeoff, never as a default for private code.
+    verified: 2026-10-07
+    source: https://opencode.ai/docs/zen
+
+  opencode/muse-spark-1.3:
+    retention: may-train
+    note: Same Meta contributor terms as the free variant.
+    verified: 2026-10-07
+    source: https://opencode.ai/docs/zen
+
+  opencode/big-pickle:
+    retention: may-train
+    note: "collected data may be used to improve the model" during its free period.
+    verified: 2026-10-07
+    source: https://opencode.ai/docs/zen
+
+  opencode/exo-free:
+    retention: may-train
+    note: "collected data may be used to improve the model" during its free period.
+    verified: 2026-10-07
+    source: https://opencode.ai/docs/zen
+
+  opencode/fledge-alpha-free:
+    retention: may-train
+    note: "collected data may be used to improve the model" during its free period.
+    verified: 2026-10-07
+    source: https://opencode.ai/docs/zen
+
+  opencode/mimo-v2.5-free:
+    retention: may-train
+    note: "collected data may be used to improve the model" during its free period.
+    verified: 2026-10-07
+    source: https://opencode.ai/docs/zen
+
+  opencode/nemotron-3-ultra-free:
+    retention: may-train
+    note: >-
+      NVIDIA free trial endpoints. Trial use only, do not submit personal or
+      confidential data; sessions are logged.
+    verified: 2026-10-07
+    source: https://opencode.ai/docs/zen

package/skills/space-build/SKILL.md

@@ -0,0 +1,54 @@
+---
+name: space-build
+description: Implement one planned task and verify it with a real command. Use when the plan is settled and files need to change.
+---
+
+# Build one task, then prove it
+
+## Steps
+
+1. **Read the task from `.space/tasks.md`, not from the conversation.** The file
+   the plan wrote is the requirement; memory of the request is not. If there is
+   no file, ask for the plan first — a task with no verify command in it is not
+   a task you can finish.
+2. **Locate before you read.** `grep`/`glob` for the candidate files, then
+   `read` with an `offset`/`limit` for the range you need. Reading whole files to
+   find out where something lives was the most expensive habit measured on this
+   setup: 2,966 full-file reads against 249 searches, about twelve per search.
+3. **Read the whole file before editing it** - not a grep hit, the file. Partial
+   reads produce edits that contradict code you did not see.
+4. **Make the smallest change** that satisfies the task. If the task is bigger
+   than it looks, stop and say so rather than quietly expanding it.
+5. **Run the verify command the plan named.** That command, not a different one,
+   not a subset.
+6. **On failure, read the actual message.** Not the last ten lines, the error.
+   Fix the cause, then run it again. The third identical attempt is a signal to
+   stop and report.
+7. **Mark the task done in `.space/tasks.md`** and paste the verify command's
+   verbatim final lines into its `verify:` line. A task marked done without
+   that has not been verified.
+8. **Report the command and its verbatim output.**
+
+## Test-first when the plan called for it
+
+If the plan said test first, write the failing test and watch it fail. A test
+that has never failed has not been shown to test anything.
+
+## Never
+
+- Leave a `TODO`, a stub, or a commented-out block for later. Partial work
+  reported as complete is the failure mode this exists to prevent.
+- Widen the task. A drive-by refactor belongs in its own plan.
+- Skip a failing test to get to green. Report it.
+- Claim a result you did not run.
+
+## Handing back
+
+Which task, which command, which result. If anything is unfinished, list it
+explicitly rather than letting the review step find it.
+
+**If you were handed back here by `space-review`**, you have a blocker with a
+`file:line` and a failing command, not a fresh task. Fix that specific thing.
+Do not re-read the plan, do not look for other improvements, and do not treat
+the review as a suggestion. A blocker that gets fixed together with a refactor
+cannot be verified on its own.