fix(channels): add 429 retry-with-backoff and Telegram send_status

[bug-ops]

Summary

• Extract Discord's existing 429 retry-with-backoff (send_with_retry, fix(channels): Discord REST client missing 429 rate-limit handling #4728) into a shared crates/zeph-channels/src/common/http_retry.rs helper and wire it into Telegram's raw API extension client (TelegramApiClient::post) and all four SlackApi methods (auth_test, post_message, update_message, download_file), which previously surfaced a rate-limited response as a hard error on the first attempt.
• Implement TelegramChannel::send_status, mirroring the existing Discord/Slack pattern, so Telegram users get the same visibility into background/implicit agent operations (skill reload, MCP elicitation, memory recall, etc.) that Discord/Slack users already see.

Closes #5949
Closes #5923

Details

• SlackApi gained a base_url field (#[cfg(test)]-only override) for wiremock testability, and its previous per-call tokio::time::timeout(15s) was replaced with a per-attempt .timeout(15s) on the RequestBuilder, since an outer timeout is incompatible with a retry loop that can legitimately run longer than one request when backing off. Worst-case wall-clock under sustained 429 is documented on the shared helper.
• TelegramChannel::send_status posts the status text as a plain-text message and no-ops when the text is empty, there is no active chat, or the channel is in a Guest Mode context (a guest reply can only be sent once via answerGuestQuery).
• Slack's Web API error messages changed shape from custom strings to reqwest::Error text (necessary to reconcile per-call timeouts with the retry loop); signatures are unchanged, and no external consumer depends on the old text (verified by grep).

Follow-ups filed during review

• no outer timeout bounds Slack/Discord channel send retries on the agent turn hot path #6094 (P1): no outer timeout bounds Slack/Discord channel send retries on the agent turn hot path (e.g. send_status in prepare_context) — pre-existing shape for Discord, now shared by Slack; worth a dedicated fix outside this PR's scope.
• Discord's send_message/edit_message/trigger_typing retry-wiring is untested end-to-end after #5949's http_retry extraction #6095 (P3): Discord's send_message/edit_message/trigger_typing retry-wiring itself isn't tested end-to-end (only the extracted helper is) — pre-existing gap, not introduced here.

Test plan

• cargo nextest run --config-file .github/nextest.toml -p zeph-channels --features "discord,slack" — 275/275 pass (34 new/changed tests: shared retry helper, Telegram retry, Slack retry + host-check-reject + debug-redaction, 4 send_status tests)
• cargo clippy --profile ci --workspace --all-targets --features "desktop,ide,server,chat,pdf,scheduler,testing" -- -D warnings — clean
• cargo nextest run --config-file .github/nextest.toml --workspace --features "desktop,ide,server,chat,pdf,scheduler" --lib --bins — 12945/12945 pass
• cargo +nightly fmt --check — clean
• RUSTFLAGS="-D warnings" RUSTDOCFLAGS="--deny rustdoc::broken_intra_doc_links" cargo doc --no-deps --workspace --features "desktop,ide,server,chat,pdf,scheduler" — clean
• RUSTFLAGS="-D warnings" cargo check --workspace --all-targets --features desktop,ide,server,chat,pdf,scheduler --locked — clean
• cargo test --doc --workspace --features "desktop,ide,server,chat,pdf,scheduler" — clean
• RUSTDOCFLAGS="--deny rustdoc::broken_intra_doc_links" cargo doc --no-deps --features "discord,slack" -p zeph-channels — clean
• CHANGELOG.md updated ([Unreleased]/Fixed)
• .local/testing/playbooks/telegram-channel.md and .local/testing/playbooks/slack.md updated with new scenarios (429 retry, send_status)
• .local/testing/coverage-status.md updated with new rows (status: Untested — not yet live-session-verified against real Telegram/Slack APIs)