Defame1297
6f6b70781d
A fresh /code-review of the APM-native authoring retarget (PR #93) found several correctness bugs beyond the ones already fixed on this branch: - new-agent.sh silently walked a marker-less subdirectory under $HOME up to user scope, contradicting its own usage text ("user scope is checked directly, no walk-up") and risking scaffolding into shared global ~/.claude or ~/.copilot directories instead of the intended local path. - The hand-copied apm.yml type: manifest detector in new-agent.sh and new-skill.sh accepted mismatched quotes (e.g. `type: "skill'`) that validate.sh's regex correctly rejects, and silently dropped a final apm.yml line lacking a trailing newline — causing the scaffolder and validator to disagree on scope for identical input. - Plugin-scope agent frontmatter could still contain the apm-agent.md template's HTML comments at ship time with no audit signal, yet apm compile copies frontmatter verbatim and <!-- --> breaks YAML parsing on both downstream harnesses. - ADR-0016 asserted agent-audit already implements a SUGGESTION heuristic for tool-restriction-needing plugin-scope agents; it doesn't. - agent-audit/README.md still described the old plugin-pair model this PR replaced with a single-file allowlist model. - validate.sh's project/user-scope CC-only/Copilot-only field checks and counterpart-missing check lost their only test coverage when the old plugin-pair fixture was deleted. Also replaces an echo-into-sed two-value parse (4 forks per call) with a single space-separated echo + read in both scaffolders. Regression tests added for every fix above, including one for a bug this pass introduced and the test suite caught: an initial two-line echo + `read` attempt silently dropped the second value, since `read` consumes only one line regardless of embedded newlines. Full suite: 158 bats tests, 39 shell-script tests, 12/12 summary categories, 0 failures. Refs: #89, #93
91 lines
5.6 KiB
Markdown
91 lines
5.6 KiB
Markdown
# Plugin-scope agent-author omits `tools:` and all Claude-only fields from `.apm/agents/*.agent.md`
|
|
|
|
This ADR is a narrower, downstream consequence discovered while designing issue #89's
|
|
implementation under ADR-0015's broader direction (Microsoft APM replaces hand-authored
|
|
plugin/marketplace authoring). It does not restate ADR-0015's rationale — see that ADR for
|
|
the parent decision.
|
|
|
|
## Context
|
|
|
|
APM's agent primitive (`.apm/agents/<name>.agent.md`) has no per-target integrator in
|
|
`apm compile` — confirmed via APM's own Python source (`integration/targets.py` and related
|
|
files, cited in `plugins/kyberforge/docs/research/docs/microsoft-apm/agent-primitive-schema.md`).
|
|
Compilation does a naive verbatim copy of the whole frontmatter and body to both the Claude
|
|
Code and Copilot CLI targets. This is unlike:
|
|
|
|
- The **skill** primitive, which is also a straight copy (confirmed in the same research doc)
|
|
but has no field semantics to conflict — `SKILL.md`'s content is target-agnostic already.
|
|
- The **prompt**, **instructions**, and **hooks** primitives, which each get real per-target
|
|
reconstruction through a dedicated integrator (field allowlisting, key renaming, dropped-field
|
|
warnings).
|
|
|
|
Because the agent primitive ships the same frontmatter unchanged to both harnesses, two
|
|
concrete incompatibilities surface:
|
|
|
|
1. **`tools:`** — Claude Code expects a space-separated tool-name string; Copilot CLI expects a
|
|
list drawn from its own alias vocabulary (`execute`/`read`/`edit`/`search`/`agent`/`web`). A
|
|
value correct for one harness is wrong for the other.
|
|
2. **Claude-only knobs with no Copilot equivalent** — `isolation`, `maxTurns`, `effort`,
|
|
`memory`, `permissionMode`. Writing any of these means Copilot's copy carries frontmatter
|
|
keys it doesn't recognize at all. Whether Copilot's agent loader ignores unknown keys or
|
|
errors on them is unconfirmed by research.
|
|
|
|
## Decision
|
|
|
|
At **plugin scope only** (destination package has an `apm.yml` at its root — an APM producer
|
|
package compiled via `apm compile`), `.apm/agents/<name>.agent.md` carries only `name`,
|
|
`description`, `model`, and the prose body. No `tools:` field, no Claude-only fields, at all.
|
|
|
|
Absent `tools:` means inherit-all-tools on both harnesses — the one value that is never wrong
|
|
on either target, unlike a present, harness-specific value that is guaranteed wrong on at least
|
|
one of them.
|
|
|
|
`agent-audit`, at plugin scope, is intended to flag — as a **SUGGESTION**, not a FAIL, since
|
|
this is an upstream schema limitation rather than an authoring mistake — any agent whose
|
|
description or body implies a need for tool restriction or a Claude-only behavior the
|
|
frontmatter can no longer express. This would give visibility into the gap without pretending
|
|
the schema can do something it can't. **Not yet implemented**: `check_apm_agent_file()` in
|
|
`validate.sh` currently validates only the field allowlist, `name`, `description`, and
|
|
body-emptiness/length — it has no heuristic for this case. Tracked as follow-up work.
|
|
|
|
### Scope boundary
|
|
|
|
This decision applies to **plugin-scope `agent-author` only**. Project scope (`.claude/agents/`
|
|
+ `.github/agents/`) and user scope (`~/.claude/agents/` + `~/.copilot/agents/`) are not APM
|
|
packages — neither goes through `apm compile` — so both keep today's dual-file Claude+Copilot
|
|
pair model exactly as ADR-0005 and ADR-0008 already describe. Those two ADRs remain fully
|
|
authoritative for project and user scope; only their plugin-scope clauses are affected by this
|
|
ADR (see the update notes appended to each).
|
|
|
|
## Considered options
|
|
|
|
**Pick one harness's vocabulary and accept breakage on the other (rejected).** E.g. always
|
|
write Claude's space-separated `tools:` string. Rejected because it ships a value that is
|
|
silently wrong (or possibly a hard error) on Copilot, and which harness "wins" would be an
|
|
arbitrary, undocumented asymmetry.
|
|
|
|
**Same as above, but `agent-audit` flags the cross-harness breakage as a tracked finding
|
|
(rejected).** Rejected for the same core reason — it still ships a wrong value to a real
|
|
harness. Tracking the breakage doesn't prevent it, and the chosen decision already gets
|
|
equivalent visibility (a SUGGESTION finding) without ever shipping the wrong value in the first
|
|
place.
|
|
|
|
## Consequences
|
|
|
|
- Every plugin-scope APM agent loses per-agent tool restriction and any Claude-only capability
|
|
(isolation, maxTurns, effort, memory, permissionMode) until APM ships a real per-target
|
|
integrator for the agent primitive. This is a known, accepted regression, not an oversight.
|
|
- **ADR-0005 is partially superseded** — its plugin-scope clause ("directory containing
|
|
`plugin.json` is plugin scope → both files land in `<root>/agents/`") no longer applies.
|
|
Plugin scope is now "directory containing `apm.yml` → single vendor-neutral file lands in
|
|
`<root>/.apm/agents/`." Project and user scope, and the rest of ADR-0005, are unaffected.
|
|
- **ADR-0008 is partially superseded** — its counterpart-derivation/pair-validation mechanism
|
|
no longer applies at plugin scope; `agent-audit` takes the single file directly there. Project
|
|
and user scope, where a real pair still exists, are unaffected.
|
|
- **ADR-0009 is not superseded.** The mechanism it established — `agent-audit` reading field
|
|
lists from `references/field-inventory.md` rather than hardcoding them, with a `source_keys`
|
|
provenance chain — survives and is reused. Only the *content shape* changes for plugin scope:
|
|
`field-inventory.md` shifts from two side-by-side CC-only/Copilot-only blocklists to one
|
|
vendor-neutral allowlist (`name`/`description`/`model`) for plugin-scope agents, while
|
|
continuing to serve its original two-blocklist role for project/user-scope validation.
|