Real work that is not worth a charter. Same item format, same gates, no epic above it. See the charter for what intake is and the eviction rule that empties it.
Filed so far: 9. Closed: 6. Three of the first five came out of E10’s convention sweep as decisions that were recorded and then not executed — which is the same failure class as a claim asserted and not checked, and is precisely the kind of work that had nowhere to live before this file existed.
| ID | Title | Priority | Effort | Status |
|---|---|---|---|---|
| BL-0058 | Read the Agent File spec; decide on a conformance line | P2 | S | ready |
| BL-0069 | Prepare repository presentation and provenance before listings | P1 | S | under-review |
| Field | Value |
|---|---|
| Epic | E00-intake |
| Pillar | P5 |
| Priority | P1 |
| Effort | S |
| Status | under-review |
| Test | partial — local checks passed; PR integration and hosted verification pending |
| Deps | — |
| Lock | — |
Why / Description: The maintainer requested all necessary housekeeping before the six proposed directory listings. Public metadata, release/version signals, license recognition, detached-skill attribution, and the existing Pages presentation must agree with the evidence.
Done means:
Plan: correct discovery metadata; fix presentation/provenance on a branch; validate the final tree and static presentation; prepare the normal PR handoff; refresh the local distribution kit. No directory submission or new methodology release is included.
Verification: recorded in E00/TEST.md when observed. Checks still pending must remain unchecked.
Maintainer direction, 2026-10-18: use the latest content for listings. The selected source is the merged v1.36.0 preview at ed65748, including v1.35.0; the older v1.34.0 tag remains historical context. This does not declare release acceptance or authorize a merge/tag/Release operation.
| Field | Value |
|---|---|
| Epic | E00-intake |
| Pillar | P4 |
| Priority | P2 |
| Effort | S |
| Status | ready |
| Test | not-tested |
| Deps | — |
| Lock | — |
Frozen intent —
Why / Description:andDone means:approved by maintainer on 2026-08-25.
Why / Description: A triage of alvinreal/awesome-opensource-ai on 2026-08-25 found the list is
overwhelmingly runtime infrastructure — inference engines, training frameworks, serving stacks — against a
methodology that governs projects using agents. Roughly nine tenths is not applicable, which is the same
result E09 got from agent-engineer and is itself the useful half of the answer.
One entry is worth reading. Agent File (.af) is an open format for serializing stateful agents with
persistent memory. Every other candidate solves state or coordination with a running service; Agent File
solves portability with a file format — the same wager this methodology makes, reached independently by
people building runtimes. It is the only entry on that list that can validate or falsify the core
premise.
The move is the BL-0034 shape, which worked: read the normative spec, compare field by field, and if the
models line up publish a compatibility signal — one sentence naming the format the handoff artifacts
resemble. A statement of fact, not a rule to learn. If they do not line up, that finding is worth the same
paragraph in FUTURE.md.
Done means:
ACTIVE_CONTEXT.md and 08’s two-layer memory,
and the comparison is recorded so it can be re-run rather than trusted.Files (probable): possibly skills/ai-dev-methodology/SKILL.md and README.md; possibly nothing.
Notes: The triage that produced this item ranked candidates from one-line descriptions and opened no repositories. That is enough to pick; it is not enough to conclude. Runners-up recorded in FUTURE.md.