coulomb/glas-harness
4
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
glas-harness/profiles/harness.agent-dev.yaml

32 lines
843 B
YAML
Raw Normal View History

Implement the harness contract and prove it against rein-aharness (GLAS-WP-0001-T02/T03/T04) - src/glas_harness/contract.py: Rein ABC (start_session/dispatch_tool/ end_session), SandboxHandle/ToolCall/ToolResult, per docs/harness-contract.md. - reins/rein_aharness.py: adapter around the rein-aharness CLI. Collapses a whole `agent-harness run` into one dispatch_tool call for now (no per-tool hooks yet — tracked in rein-aharness HARNESS-WP-0002-T03); verifies success via new-commit detection, mirroring rein-aharness's own success signal. - gateway.py + cli.py: resolves a sand-boxer sandbox, runs one task through a rein, tears the sandbox down. - registry/reins/*.yaml (rein-aharness implemented, rein-openweights planned) and profiles/harness.agent-dev{,-local}.yaml, pairing with sand-boxer's profile.agent-dev and the new profile.bwrap-local. Tested against a real local git repo + mocked SandboxManager/CLI subprocess (10 tests, all passing). A real live run against the actual Claude Code CLI is deliberately left for a human-triggered follow-up — not executed autonomously since it spends real API credits/credentials. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
2026-07-26 13:02:59 +02:00
id: harness.agent-dev
version: "1.0.0"
feat: add versioned execution profiles
2026-08-21 00:21:53 +02:00
contract_version: "1.0"
status: enabled
Make profile runtime readiness explicit Assistant: codex Assistant-Model: gpt-5.6-sol Assistant-Session: 01a0233b-178d-7162-b92f-31a31ea8ca9b
2026-08-23 11:32:23 +02:00
operational_readiness:
status: unverified
reason: no positive remote execution proof exists after transport hardening
owner: glas-harness
evidence_ref: GLAS-WP-0005
feat: add versioned execution profiles
2026-08-21 00:21:53 +02:00
rein:
id: rein-aharness
required_capabilities:
session_style: unattended
model_class: frontier
Implement the harness contract and prove it against rein-aharness (GLAS-WP-0001-T02/T03/T04) - src/glas_harness/contract.py: Rein ABC (start_session/dispatch_tool/ end_session), SandboxHandle/ToolCall/ToolResult, per docs/harness-contract.md. - reins/rein_aharness.py: adapter around the rein-aharness CLI. Collapses a whole `agent-harness run` into one dispatch_tool call for now (no per-tool hooks yet — tracked in rein-aharness HARNESS-WP-0002-T03); verifies success via new-commit detection, mirroring rein-aharness's own success signal. - gateway.py + cli.py: resolves a sand-boxer sandbox, runs one task through a rein, tears the sandbox down. - registry/reins/*.yaml (rein-aharness implemented, rein-openweights planned) and profiles/harness.agent-dev{,-local}.yaml, pairing with sand-boxer's profile.agent-dev and the new profile.bwrap-local. Tested against a real local git repo + mocked SandboxManager/CLI subprocess (10 tests, all passing). A real live run against the actual Claude Code CLI is deliberately left for a human-triggered follow-up — not executed autonomously since it spends real API credits/credentials. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
2026-07-26 13:02:59 +02:00
sandbox_profile: profile.agent-dev
tool_profile: green-commit-only
feat: add versioned execution profiles
2026-08-21 00:21:53 +02:00
model:
provider: anthropic
model: claude-sonnet-4-6
model_class: frontier
route: claude-code-cli
limits:
budget_tokens: 60000
timeout_seconds: 900
credential_route_refs: []
Implement the harness contract and prove it against rein-aharness (GLAS-WP-0001-T02/T03/T04) - src/glas_harness/contract.py: Rein ABC (start_session/dispatch_tool/ end_session), SandboxHandle/ToolCall/ToolResult, per docs/harness-contract.md. - reins/rein_aharness.py: adapter around the rein-aharness CLI. Collapses a whole `agent-harness run` into one dispatch_tool call for now (no per-tool hooks yet — tracked in rein-aharness HARNESS-WP-0002-T03); verifies success via new-commit detection, mirroring rein-aharness's own success signal. - gateway.py + cli.py: resolves a sand-boxer sandbox, runs one task through a rein, tears the sandbox down. - registry/reins/*.yaml (rein-aharness implemented, rein-openweights planned) and profiles/harness.agent-dev{,-local}.yaml, pairing with sand-boxer's profile.agent-dev and the new profile.bwrap-local. Tested against a real local git repo + mocked SandboxManager/CLI subprocess (10 tests, all passing). A real live run against the actual Claude Code CLI is deliberately left for a human-triggered follow-up — not executed autonomously since it spends real API credits/credentials. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
2026-07-26 13:02:59 +02:00
metadata:
latency_class: standard
feat: add versioned execution profiles
2026-08-21 00:21:53 +02:00
stream_tool_events: true
Implement the harness contract and prove it against rein-aharness (GLAS-WP-0001-T02/T03/T04) - src/glas_harness/contract.py: Rein ABC (start_session/dispatch_tool/ end_session), SandboxHandle/ToolCall/ToolResult, per docs/harness-contract.md. - reins/rein_aharness.py: adapter around the rein-aharness CLI. Collapses a whole `agent-harness run` into one dispatch_tool call for now (no per-tool hooks yet — tracked in rein-aharness HARNESS-WP-0002-T03); verifies success via new-commit detection, mirroring rein-aharness's own success signal. - gateway.py + cli.py: resolves a sand-boxer sandbox, runs one task through a rein, tears the sandbox down. - registry/reins/*.yaml (rein-aharness implemented, rein-openweights planned) and profiles/harness.agent-dev{,-local}.yaml, pairing with sand-boxer's profile.agent-dev and the new profile.bwrap-local. Tested against a real local git repo + mocked SandboxManager/CLI subprocess (10 tests, all passing). A real live run against the actual Claude Code CLI is deliberately left for a human-triggered follow-up — not executed autonomously since it spends real API credits/credentials. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
2026-07-26 13:02:59 +02:00
# Pairs with sand-boxer's profile.agent-dev (ext.compose-ssh, remote
# host over SSH). Use harness.agent-dev-local for fast local iteration
# without the SSH hop.
Reference in a new issue Copy permalink