disinto-admin/disinto

Fork 0

fix: Two parallel activation paths for llama agents (ENABLE_LLAMA_AGENT vs [agents.X] TOML) (#846) #906

Merged

dev-qwen2 merged 2 commits from fix/issue-846 into main

2026-04-16 19:22:44 +00:00

dev-qwen2 commented

2026-04-16 18:42:21 +00:00

Collaborator

Fixes #846

Changes

Fixes #846 ## Changes

dev-qwen2 added 1 commit 2026-04-16 18:42:21 +00:00

fix: Two parallel activation paths for llama agents (ENABLE_LLAMA_AGENT vs [agents.X] TOML) (#846 )

ci/woodpecker/push/ci Pipeline was successful

Details

ci/woodpecker/push/nomad-validate Pipeline was successful

Details

ci/woodpecker/pr/ci Pipeline was successful

Details

ci/woodpecker/pr/nomad-validate Pipeline was successful

Details

ci/woodpecker/pr/secret-scan Pipeline was successful

Details

ci/woodpecker/pr/smoke-init Pipeline was successful

Details

11f89d9b7a

review-bot commented

2026-04-16 18:54:47 +00:00

Collaborator

AI Review

Summary

The PR correctly removes the legacy ENABLE_LLAMA_AGENT activation path: migration guard in entrypoint, env var cleanup, forge-setup pruning, and generator removal all look correct. Two blocking omissions prevent merge.

Blocking

high AGENTS.md:125-126: Documents agents-llama and agents-llama-all as gated on ENABLE_LLAMA_AGENT=1 — the exact mechanism this PR removes. Per the review contract (step 3b), a behavioral change that contradicts documented behavior requires a doc update in the same PR. Update these two rows to describe the TOML [agents.X] path, or remove them and note that llama agents are provisioned via hire-an-agent.
high formulas/release.sh:181-182 / formulas/release.toml:192-195,206: Both files still reference agents-llama as a named compose service. After a user runs disinto init post-merge, the regenerated docker-compose.yml has no agents-llama service. The next release formula invocation will fail at the up command because compose returns a non-zero exit when a named service is not defined. Remove agents-llama from both commands (TOML-path llama agents are named agents-dev-qwen etc., not agents-llama).

Non-blocking

low tests/fixtures/dot-env-complete:28-29 / dot-env-incomplete:15-16: Fixtures still carry FORGE_TOKEN_LLAMA / FORGE_PASS_LLAMA. tools/vault-import.sh still consumes them so existing vault-import tests pass, but the fixture represents a credentials model that no longer exists. Update when vault-import.sh is cleaned up.
low docker-compose.yml:67-134: The committed file still has the agents-llama and agents-llama-all stanzas using FORGE_TOKEN_LLAMA/FORGE_PASS_LLAMA. This is a generated artifact; disinto init will overwrite it. Regenerate and commit the updated file so the checked-in state matches the new generator behavior.

Verdict

REQUEST_CHANGES — AGENTS.md documents removed ENABLE_LLAMA_AGENT mechanism; release formula references deleted agents-llama service causing post-migration breakage

Reviewed at 11f89d9 | AGENTS.md

## AI Review  ### Summary The PR correctly removes the legacy `ENABLE_LLAMA_AGENT` activation path: migration guard in entrypoint, env var cleanup, forge-setup pruning, and generator removal all look correct. Two blocking omissions prevent merge. --- ### Blocking - **high** `AGENTS.md:125-126`: Documents `agents-llama` and `agents-llama-all` as gated on `ENABLE_LLAMA_AGENT=1` — the exact mechanism this PR removes. Per the review contract (step 3b), a behavioral change that contradicts documented behavior requires a doc update in the same PR. Update these two rows to describe the TOML `[agents.X]` path, or remove them and note that llama agents are provisioned via `hire-an-agent`. - **high** `formulas/release.sh:181-182` / `formulas/release.toml:192-195,206`: Both files still reference `agents-llama` as a named compose service. After a user runs `disinto init` post-merge, the regenerated `docker-compose.yml` has no `agents-llama` service. The next release formula invocation will fail at the `up` command because compose returns a non-zero exit when a named service is not defined. Remove `agents-llama` from both commands (TOML-path llama agents are named `agents-dev-qwen` etc., not `agents-llama`). --- ### Non-blocking - **low** `tests/fixtures/dot-env-complete:28-29` / `dot-env-incomplete:15-16`: Fixtures still carry `FORGE_TOKEN_LLAMA` / `FORGE_PASS_LLAMA`. `tools/vault-import.sh` still consumes them so existing vault-import tests pass, but the fixture represents a credentials model that no longer exists. Update when vault-import.sh is cleaned up. - **low** `docker-compose.yml:67-134`: The committed file still has the `agents-llama` and `agents-llama-all` stanzas using `FORGE_TOKEN_LLAMA`/`FORGE_PASS_LLAMA`. This is a generated artifact; `disinto init` will overwrite it. Regenerate and commit the updated file so the checked-in state matches the new generator behavior. ### Verdict **REQUEST_CHANGES** — AGENTS.md documents removed ENABLE_LLAMA_AGENT mechanism; release formula references deleted agents-llama service causing post-migration breakage --- *Reviewed at `11f89d9` | [AGENTS.md](AGENTS.md)*

review-bot requested changes 2026-04-16 18:54:47 +00:00

Dismissed

review-bot left a comment

Collaborator

AI Review: REQUEST_CHANGES — AGENTS.md documents removed ENABLE_LLAMA_AGENT mechanism; release formula references deleted agents-llama service causing post-migration breakage

AI Review: **REQUEST_CHANGES** — AGENTS.md documents removed ENABLE_LLAMA_AGENT mechanism; release formula references deleted agents-llama service causing post-migration breakage

dev-qwen2 force-pushed fix/issue-846 from 11f89d9b7a to e003829eaa

2026-04-16 19:05:51 +00:00

Compare

review-bot commented

2026-04-16 19:22:27 +00:00

Collaborator

AI Re-review (round 2)

Previous Findings\n- AGENTS.md rows documenting ENABLE_LLAMA_AGENT gate → FIXED: replaced with single "Local-model agents" row pointing to `[agents.X]` TOML\n- formulas/release.sh + release.toml agents-llama references → FIXED: `agents-llama` removed from `stop`, `up`, `down`, and `ps` commands in both files\n\n### New Issues\n\nNone. The `entrypoint.sh` hunk in the incremental diff is the already-merged #877 fix (`c77fb1d`), not a new change from this branch. No collateral damage in the fixup commits.

Verdict

APPROVE — Both blockers closed: AGENTS.md updated to TOML path, release formula stripped of agents-llama references

Reviewed at e003829 | Previous: 11f89d9 | AGENTS.md

## AI Re-review (round 2)  ### Previous Findings\n- **AGENTS.md rows documenting ENABLE_LLAMA_AGENT gate** → FIXED: replaced with single "Local-model agents" row pointing to `[agents.X]` TOML\n- **formulas/release.sh + release.toml agents-llama references** → FIXED: `agents-llama` removed from `stop`, `up`, `down`, and `ps` commands in both files\n\n### New Issues\n\nNone. The `entrypoint.sh` hunk in the incremental diff is the already-merged #877 fix (c77fb1d), not a new change from this branch. No collateral damage in the fixup commits. ### Verdict **APPROVE** — Both blockers closed: AGENTS.md updated to TOML path, release formula stripped of agents-llama references --- *Reviewed at `e003829` | Previous: `11f89d9` | [AGENTS.md](AGENTS.md)*

review-bot approved these changes 2026-04-16 19:22:27 +00:00

review-bot left a comment

Collaborator

AI Re-review (round 2): APPROVE — Both blockers closed: AGENTS.md updated to TOML path, release formula stripped of agents-llama references

AI Re-review (round 2): **APPROVE** — Both blockers closed: AGENTS.md updated to TOML path, release formula stripped of agents-llama references