fix: bug: hire-an-agent does not add the new agent as collaborator on the project repo (#856)
All checks were successful
ci/woodpecker/push/ci Pipeline was successful
ci/woodpecker/pr/ci Pipeline was successful
ci/woodpecker/pr/smoke-init Pipeline was successful

hire-an-agent now adds the new Forgejo user as a `write` collaborator on
`$FORGE_REPO` right after the token step, mirroring the collaborator setup
lib/forge-setup.sh applies to the canonical bot users. Without this, a
freshly hired agent's PATCH to assign itself an issue returned 403 Forbidden
and the dev-agent polled forever logging "claim lost to <none>".

issue_claim() now captures the PATCH HTTP status via `-w '%{http_code}'`
instead of swallowing failures with `curl -sf ... || return 1`. A 403 (or
any non-2xx) now surfaces a distinct log line naming the code — the missing
collaborator root cause would have been diagnosable in seconds instead of
minutes.

Also updates the lib-issue-claim bats mock to handle the new `-w` flag and
adds a regression test covering the HTTP-error log surfacing path.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
This commit is contained in:
Claude 2026-04-16 10:47:51 +00:00
parent 120bce5745
commit 9d5cbb4fa2
3 changed files with 74 additions and 4 deletions

View file

@ -229,6 +229,37 @@ disinto_hire_an_agent() {
export "${pass_var}=${user_pass}"
fi
# Step 1.6: Add the new agent as a write collaborator on the project repo (#856).
# Without this, PATCH /issues/{n} {assignees:[agent]} returns 403 Forbidden and
# the dev-agent polls forever logging "claim lost to <none> — skipping" (see
# issue_claim()'s post-PATCH verify). Mirrors the collaborator setup applied
# to the canonical bot users in lib/forge-setup.sh. Idempotent: Forgejo's PUT
# returns 204 whether the user is being added for the first time or already a
# collaborator at the same permission.
if [ -n "${FORGE_REPO:-}" ]; then
echo ""
echo "Step 1.6: Adding '${agent_name}' as write collaborator on '${FORGE_REPO}'..."
local collab_code
collab_code=$(curl -s -o /dev/null -w '%{http_code}' -X PUT \
-H "Authorization: token ${admin_token}" \
-H "Content-Type: application/json" \
"${forge_url}/api/v1/repos/${FORGE_REPO}/collaborators/${agent_name}" \
-d '{"permission":"write"}')
case "$collab_code" in
204|201|200)
echo " ${agent_name} is a write collaborator on ${FORGE_REPO} (HTTP ${collab_code})"
;;
*)
echo " Warning: failed to add '${agent_name}' as collaborator on '${FORGE_REPO}' (HTTP ${collab_code})" >&2
echo " The agent will not be able to claim issues until this is fixed." >&2
;;
esac
else
echo ""
echo "Step 1.6: FORGE_REPO not set — skipping collaborator step" >&2
echo " Warning: the agent will not be able to claim issues on the project repo" >&2
fi
# Step 2: Create .profile repo on Forgejo
echo ""
echo "Step 2: Creating '${agent_name}/.profile' repo (if not exists)..."

View file

@ -126,11 +126,21 @@ issue_claim() {
# Assign to self BEFORE adding in-progress label (issue #471).
# This ordering ensures the assignee is set by the time other pollers
# see the in-progress label, reducing the stale-detection race window.
curl -sf -X PATCH \
#
# Capture the HTTP status instead of silently swallowing failures (#856).
# A 403 here means the bot user is not a write collaborator on the repo —
# previously the silent failure fell through to the post-PATCH verify which
# only reported "claim lost to <none>", hiding the real root cause.
local patch_code
patch_code=$(curl -s -o /dev/null -w '%{http_code}' -X PATCH \
-H "Authorization: token ${FORGE_TOKEN}" \
-H "Content-Type: application/json" \
"${FORGE_API}/issues/${issue}" \
-d "{\"assignees\":[\"${me}\"]}" >/dev/null 2>&1 || return 1
-d "{\"assignees\":[\"${me}\"]}")
if [ "$patch_code" != "201" ] && [ "$patch_code" != "200" ]; then
_ilc_log "issue #${issue} PATCH assignee failed: HTTP ${patch_code} (403 = missing write collaborator permission on ${FORGE_REPO:-repo})"
return 1
fi
# Verify the PATCH stuck. Forgejo's assignees PATCH is last-write-wins, so
# under concurrent claims from multiple dev agents two invocations can both

View file

@ -52,12 +52,13 @@ setup() {
# canned responses per endpoint. Every call gets logged as
# `METHOD URL` (one line) to $CALLS_LOG for later grep-based asserts.
curl() {
local method="GET" url="" arg
local method="GET" url="" arg want_code=""
while [ $# -gt 0 ]; do
arg="$1"
case "$arg" in
-X) method="$2"; shift 2 ;;
-H|-d|--data-binary|-o) shift 2 ;;
-w) want_code="$2"; shift 2 ;;
-sf|-s|-f|--silent|--fail) shift ;;
*) url="$arg"; shift ;;
esac
@ -89,7 +90,13 @@ setup() {
fi
;;
"PATCH ${FORGE_API}/issues/"*)
: # accept any PATCH; body is ignored by the mock
# Accept any PATCH; body ignored. When caller asked for the HTTP
# status via `-w '%{http_code}'` (issue_claim does this since #856
# to surface 403s from missing collaborator permission), emit the
# code configured by the scenario (default 200).
if [ "$want_code" = '%{http_code}' ]; then
printf '%s' "${MOCK_PATCH_CODE:-200}"
fi
;;
"GET ${FORGE_API}/labels")
printf '[]'
@ -165,6 +172,28 @@ count_calls() {
[ "$(count_calls GET "${FORGE_API}/labels")" -eq 0 ]
}
# ── PATCH HTTP error surfacing (#856) ───────────────────────────────────────
@test "issue_claim logs specific HTTP code on PATCH failure (403 = missing collaborator)" {
export MOCK_ME="bot"
export MOCK_INITIAL_ASSIGNEE=""
export MOCK_RECHECK_ASSIGNEE=""
export MOCK_PATCH_CODE="403"
run issue_claim 42
[ "$status" -eq 1 ]
# The new log message names the HTTP code explicitly — without this,
# a missing-collaborator setup (#856) falls through to the post-PATCH
# verify and masquerades as "claim lost to <none>".
[[ "$output" == *"PATCH assignee failed: HTTP 403"* ]]
# No re-read on PATCH failure (we bail before reaching the verify step).
[ "$(count_calls GET "${FORGE_API}/issues/42")" -eq 1 ]
[ "$(count_calls PATCH "${FORGE_API}/issues/42")" -eq 1 ]
[ "$(count_calls POST "${FORGE_API}/issues/42/labels")" -eq 0 ]
}
# ── pre-check skip ──────────────────────────────────────────────────────────
@test "issue_claim skips early (no PATCH) when pre-check shows another assignee" {