Call the gitea-tools MCP directly instead of the parked gitea-agent #11

2026-08-18T16:06:47+03:00

oleks commented

2026-08-18 16:06:47 +03:00

Closes #10.

Six sites across five files. cluster:gitea-agent was parked in cluster v0.87.0; anxious:issuer-agent in anxious v0.64.0.

Site	Change
`skills/retry-tag-hygiene/SKILL.md:170-171`	`list_tags` / `delete_tag` now named on the `gitea-tools` MCP. This was the skill's actual mechanism, so it could not complete.
`skills/retry-tag-hygiene/references/group-retry-tags.py:11`	docstring names the MCP, not the agent
`skills/woodpecker-trigger-discipline/SKILL.md:366`	reworded to "the since-parked gitea agent" — the incident is real history and worth keeping; the `plugin:component` form is what dangles
`skills/woodpecker-trigger-discipline/SKILL.md:372`	delegation-preference list now leads with the `gitea-tools` MCP, which is where `branch_protection_read` actually lives
`specs/001-.../quickstart.md:62`	acceptance assertion no longer expects delegation naming the dead agent — it would have re-pinned the old path
`HANDOFF.md:243`	bug filing routed to `delivery-tools` (`prepare-filing`, then `file-issue`)

Note the 372 row is a small correctness improvement beyond the rename: the old text justified the delegation as "cluster:gitea-agent owns git.oleks.space reads, has tea api", but the tool the surrounding section actually needs — branch_protection_read — is a plain gitea-tools verb, so no delegation is required at all.

FACTBOOK-woodpecker-hygiene.md:248 is deliberately untouched: it is a past-tense incident record, not an instruction.

fleet-govern check-ownership now reports zero blocking findings for this repo (was 4).

Version bumped 1.4.9 → 1.4.10, tagged v1.4.10.

Closes #10. Six sites across five files. `cluster:gitea-agent` was parked in cluster v0.87.0; `anxious:issuer-agent` in anxious v0.64.0. | Site | Change | |---|---| | `skills/retry-tag-hygiene/SKILL.md:170-171` | `list_tags` / `delete_tag` now named on the `gitea-tools` MCP. **This was the skill's actual mechanism**, so it could not complete. | | `skills/retry-tag-hygiene/references/group-retry-tags.py:11` | docstring names the MCP, not the agent | | `skills/woodpecker-trigger-discipline/SKILL.md:366` | reworded to "the since-parked gitea agent" — the incident is real history and worth keeping; the `plugin:component` form is what dangles | | `skills/woodpecker-trigger-discipline/SKILL.md:372` | delegation-preference list now leads with the `gitea-tools` MCP, which is where `branch_protection_read` actually lives | | `specs/001-.../quickstart.md:62` | acceptance assertion no longer expects delegation *naming the dead agent* — it would have re-pinned the old path | | `HANDOFF.md:243` | bug filing routed to `delivery-tools` (`prepare-filing`, then `file-issue`) | Note the 372 row is a small correctness improvement beyond the rename: the old text justified the delegation as "`cluster:gitea-agent` owns git.oleks.space reads, has `tea api`", but the tool the surrounding section actually needs — `branch_protection_read` — is a plain `gitea-tools` verb, so no delegation is required at all. `FACTBOOK-woodpecker-hygiene.md:248` is deliberately untouched: it is a past-tense incident record, not an instruction. **`fleet-govern check-ownership` now reports zero blocking findings for this repo** (was 4). Version bumped 1.4.9 → 1.4.10, tagged `v1.4.10`.

oleks added 1 commit 2026-08-18 16:06:47 +03:00

call the gitea-tools MCP directly instead of the parked gitea-agent (oleks/claude-plugin-cicd-insights#10 ), v1.4.10 8587121ee9

oleks merged commit 08763fe186 into main

2026-08-18 16:06:51 +03:00

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: oleks/claude-plugin-cicd-insights#11